From 71cf44d51fd7306ea0ebcce44b87e79b4fbe2da1 Mon Sep 17 00:00:00 2001
From: ModelHub XC <noreply@modelhub.org.cn>
Date: Mon, 20 Jul 2026 00:47:07 +0800
Subject: [PATCH] =?UTF-8?q?=E5=88=9D=E5=A7=8B=E5=8C=96=E9=A1=B9=E7=9B=AE?=
 =?UTF-8?q?=EF=BC=8C=E7=94=B1ModelHub=20XC=E7=A4=BE=E5=8C=BA=E6=8F=90?=
 =?UTF-8?q?=E4=BE=9B=E6=A8=A1=E5=9E=8B?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Model: shisa-ai/ablation-145-a128.dpo.armorm.rp.tl.8e7-shisa-v2-llama-3.1-8b
Source: Original Platform
---
 .gitattributes                                |    52 +
 README.md                                     |    66 +
 checkpoint-1630/config.json                   |    35 +
 checkpoint-1630/generation_config.json        |    12 +
 ..._zero_pp_rank_0_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_10_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_11_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_12_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_13_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_14_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_15_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_16_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_17_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_18_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_19_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_1_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_20_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_21_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_22_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_23_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_24_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_25_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_26_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_27_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_28_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_29_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_2_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_30_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_31_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_3_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_4_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_5_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_6_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_7_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_8_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_9_mp_rank_00_optim_states.pt |     3 +
 .../zero_pp_rank_0_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_10_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_11_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_12_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_13_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_14_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_15_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_16_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_17_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_18_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_19_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_1_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_20_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_21_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_22_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_23_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_24_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_25_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_26_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_27_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_28_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_29_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_2_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_30_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_31_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_3_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_4_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_5_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_6_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_7_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_8_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_9_mp_rank_00_model_states.pt |     3 +
 checkpoint-1630/latest                        |     1 +
 .../model-00001-of-00004.safetensors          |     3 +
 .../model-00002-of-00004.safetensors          |     3 +
 .../model-00003-of-00004.safetensors          |     3 +
 .../model-00004-of-00004.safetensors          |     3 +
 checkpoint-1630/model.safetensors.index.json  |   298 +
 checkpoint-1630/rng_state_0.pth               |     3 +
 checkpoint-1630/rng_state_1.pth               |     3 +
 checkpoint-1630/rng_state_10.pth              |     3 +
 checkpoint-1630/rng_state_11.pth              |     3 +
 checkpoint-1630/rng_state_12.pth              |     3 +
 checkpoint-1630/rng_state_13.pth              |     3 +
 checkpoint-1630/rng_state_14.pth              |     3 +
 checkpoint-1630/rng_state_15.pth              |     3 +
 checkpoint-1630/rng_state_16.pth              |     3 +
 checkpoint-1630/rng_state_17.pth              |     3 +
 checkpoint-1630/rng_state_18.pth              |     3 +
 checkpoint-1630/rng_state_19.pth              |     3 +
 checkpoint-1630/rng_state_2.pth               |     3 +
 checkpoint-1630/rng_state_20.pth              |     3 +
 checkpoint-1630/rng_state_21.pth              |     3 +
 checkpoint-1630/rng_state_22.pth              |     3 +
 checkpoint-1630/rng_state_23.pth              |     3 +
 checkpoint-1630/rng_state_24.pth              |     3 +
 checkpoint-1630/rng_state_25.pth              |     3 +
 checkpoint-1630/rng_state_26.pth              |     3 +
 checkpoint-1630/rng_state_27.pth              |     3 +
 checkpoint-1630/rng_state_28.pth              |     3 +
 checkpoint-1630/rng_state_29.pth              |     3 +
 checkpoint-1630/rng_state_3.pth               |     3 +
 checkpoint-1630/rng_state_30.pth              |     3 +
 checkpoint-1630/rng_state_31.pth              |     3 +
 checkpoint-1630/rng_state_4.pth               |     3 +
 checkpoint-1630/rng_state_5.pth               |     3 +
 checkpoint-1630/rng_state_6.pth               |     3 +
 checkpoint-1630/rng_state_7.pth               |     3 +
 checkpoint-1630/rng_state_8.pth               |     3 +
 checkpoint-1630/rng_state_9.pth               |     3 +
 checkpoint-1630/scheduler.pt                  |     3 +
 checkpoint-1630/special_tokens_map.json       |    23 +
 checkpoint-1630/tokenizer.json                |     3 +
 checkpoint-1630/tokenizer_config.json         |  2064 +
 checkpoint-1630/trainer_state.json            | 24484 +++++++++++
 checkpoint-1630/training_args.bin             |     3 +
 checkpoint-1630/zero_to_fp32.py               |   604 +
 checkpoint-2445/config.json                   |    35 +
 checkpoint-2445/generation_config.json        |    12 +
 ..._zero_pp_rank_0_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_10_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_11_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_12_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_13_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_14_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_15_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_16_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_17_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_18_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_19_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_1_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_20_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_21_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_22_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_23_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_24_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_25_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_26_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_27_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_28_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_29_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_2_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_30_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_31_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_3_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_4_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_5_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_6_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_7_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_8_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_9_mp_rank_00_optim_states.pt |     3 +
 .../zero_pp_rank_0_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_10_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_11_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_12_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_13_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_14_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_15_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_16_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_17_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_18_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_19_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_1_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_20_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_21_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_22_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_23_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_24_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_25_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_26_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_27_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_28_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_29_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_2_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_30_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_31_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_3_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_4_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_5_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_6_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_7_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_8_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_9_mp_rank_00_model_states.pt |     3 +
 checkpoint-2445/latest                        |     1 +
 .../model-00001-of-00004.safetensors          |     3 +
 .../model-00002-of-00004.safetensors          |     3 +
 .../model-00003-of-00004.safetensors          |     3 +
 .../model-00004-of-00004.safetensors          |     3 +
 checkpoint-2445/model.safetensors.index.json  |   298 +
 checkpoint-2445/rng_state_0.pth               |     3 +
 checkpoint-2445/rng_state_1.pth               |     3 +
 checkpoint-2445/rng_state_10.pth              |     3 +
 checkpoint-2445/rng_state_11.pth              |     3 +
 checkpoint-2445/rng_state_12.pth              |     3 +
 checkpoint-2445/rng_state_13.pth              |     3 +
 checkpoint-2445/rng_state_14.pth              |     3 +
 checkpoint-2445/rng_state_15.pth              |     3 +
 checkpoint-2445/rng_state_16.pth              |     3 +
 checkpoint-2445/rng_state_17.pth              |     3 +
 checkpoint-2445/rng_state_18.pth              |     3 +
 checkpoint-2445/rng_state_19.pth              |     3 +
 checkpoint-2445/rng_state_2.pth               |     3 +
 checkpoint-2445/rng_state_20.pth              |     3 +
 checkpoint-2445/rng_state_21.pth              |     3 +
 checkpoint-2445/rng_state_22.pth              |     3 +
 checkpoint-2445/rng_state_23.pth              |     3 +
 checkpoint-2445/rng_state_24.pth              |     3 +
 checkpoint-2445/rng_state_25.pth              |     3 +
 checkpoint-2445/rng_state_26.pth              |     3 +
 checkpoint-2445/rng_state_27.pth              |     3 +
 checkpoint-2445/rng_state_28.pth              |     3 +
 checkpoint-2445/rng_state_29.pth              |     3 +
 checkpoint-2445/rng_state_3.pth               |     3 +
 checkpoint-2445/rng_state_30.pth              |     3 +
 checkpoint-2445/rng_state_31.pth              |     3 +
 checkpoint-2445/rng_state_4.pth               |     3 +
 checkpoint-2445/rng_state_5.pth               |     3 +
 checkpoint-2445/rng_state_6.pth               |     3 +
 checkpoint-2445/rng_state_7.pth               |     3 +
 checkpoint-2445/rng_state_8.pth               |     3 +
 checkpoint-2445/rng_state_9.pth               |     3 +
 checkpoint-2445/scheduler.pt                  |     3 +
 checkpoint-2445/special_tokens_map.json       |    23 +
 checkpoint-2445/tokenizer.json                |     3 +
 checkpoint-2445/tokenizer_config.json         |  2064 +
 checkpoint-2445/trainer_state.json            | 36709 ++++++++++++++++
 checkpoint-2445/training_args.bin             |     3 +
 checkpoint-2445/zero_to_fp32.py               |   604 +
 checkpoint-815/config.json                    |    35 +
 checkpoint-815/generation_config.json         |    12 +
 ..._zero_pp_rank_0_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_10_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_11_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_12_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_13_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_14_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_15_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_16_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_17_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_18_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_19_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_1_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_20_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_21_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_22_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_23_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_24_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_25_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_26_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_27_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_28_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_29_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_2_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_30_mp_rank_00_optim_states.pt |     3 +
 ...zero_pp_rank_31_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_3_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_4_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_5_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_6_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_7_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_8_mp_rank_00_optim_states.pt |     3 +
 ..._zero_pp_rank_9_mp_rank_00_optim_states.pt |     3 +
 .../zero_pp_rank_0_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_10_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_11_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_12_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_13_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_14_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_15_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_16_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_17_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_18_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_19_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_1_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_20_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_21_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_22_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_23_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_24_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_25_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_26_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_27_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_28_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_29_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_2_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_30_mp_rank_00_model_states.pt |     3 +
 ...zero_pp_rank_31_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_3_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_4_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_5_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_6_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_7_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_8_mp_rank_00_model_states.pt |     3 +
 .../zero_pp_rank_9_mp_rank_00_model_states.pt |     3 +
 checkpoint-815/latest                         |     1 +
 .../model-00001-of-00004.safetensors          |     3 +
 .../model-00002-of-00004.safetensors          |     3 +
 .../model-00003-of-00004.safetensors          |     3 +
 .../model-00004-of-00004.safetensors          |     3 +
 checkpoint-815/model.safetensors.index.json   |   298 +
 checkpoint-815/rng_state_0.pth                |     3 +
 checkpoint-815/rng_state_1.pth                |     3 +
 checkpoint-815/rng_state_10.pth               |     3 +
 checkpoint-815/rng_state_11.pth               |     3 +
 checkpoint-815/rng_state_12.pth               |     3 +
 checkpoint-815/rng_state_13.pth               |     3 +
 checkpoint-815/rng_state_14.pth               |     3 +
 checkpoint-815/rng_state_15.pth               |     3 +
 checkpoint-815/rng_state_16.pth               |     3 +
 checkpoint-815/rng_state_17.pth               |     3 +
 checkpoint-815/rng_state_18.pth               |     3 +
 checkpoint-815/rng_state_19.pth               |     3 +
 checkpoint-815/rng_state_2.pth                |     3 +
 checkpoint-815/rng_state_20.pth               |     3 +
 checkpoint-815/rng_state_21.pth               |     3 +
 checkpoint-815/rng_state_22.pth               |     3 +
 checkpoint-815/rng_state_23.pth               |     3 +
 checkpoint-815/rng_state_24.pth               |     3 +
 checkpoint-815/rng_state_25.pth               |     3 +
 checkpoint-815/rng_state_26.pth               |     3 +
 checkpoint-815/rng_state_27.pth               |     3 +
 checkpoint-815/rng_state_28.pth               |     3 +
 checkpoint-815/rng_state_29.pth               |     3 +
 checkpoint-815/rng_state_3.pth                |     3 +
 checkpoint-815/rng_state_30.pth               |     3 +
 checkpoint-815/rng_state_31.pth               |     3 +
 checkpoint-815/rng_state_4.pth                |     3 +
 checkpoint-815/rng_state_5.pth                |     3 +
 checkpoint-815/rng_state_6.pth                |     3 +
 checkpoint-815/rng_state_7.pth                |     3 +
 checkpoint-815/rng_state_8.pth                |     3 +
 checkpoint-815/rng_state_9.pth                |     3 +
 checkpoint-815/scheduler.pt                   |     3 +
 checkpoint-815/special_tokens_map.json        |    23 +
 checkpoint-815/tokenizer.json                 |     3 +
 checkpoint-815/tokenizer_config.json          |  2064 +
 checkpoint-815/trainer_state.json             | 12259 ++++++
 checkpoint-815/training_args.bin              |     3 +
 checkpoint-815/zero_to_fp32.py                |   604 +
 config.json                                   |    35 +
 configuration.json                            |     1 +
 generation_config.json                        |    12 +
 model-00001-of-00004.safetensors              |     3 +
 model-00002-of-00004.safetensors              |     3 +
 model-00003-of-00004.safetensors              |     3 +
 model-00004-of-00004.safetensors              |     3 +
 model.safetensors.index.json                  |   298 +
 ...fevents.1743660284.ip-10-1-21-143.961514.0 |     3 +
 special_tokens_map.json                       |    23 +
 tokenizer.json                                |     3 +
 tokenizer_config.json                         |  2064 +
 training_args.bin                             |     3 +
 348 files changed, 86062 insertions(+)
 create mode 100644 .gitattributes
 create mode 100644 README.md
 create mode 100644 checkpoint-1630/config.json
 create mode 100644 checkpoint-1630/generation_config.json
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_0_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_10_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_11_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_12_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_13_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_14_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_15_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_16_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_17_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_18_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_19_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_1_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_20_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_21_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_22_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_23_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_24_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_25_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_26_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_27_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_28_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_29_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_2_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_30_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_31_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_3_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_4_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_5_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_6_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_7_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_8_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/global_step1630/zero_pp_rank_9_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-1630/latest
 create mode 100644 checkpoint-1630/model-00001-of-00004.safetensors
 create mode 100644 checkpoint-1630/model-00002-of-00004.safetensors
 create mode 100644 checkpoint-1630/model-00003-of-00004.safetensors
 create mode 100644 checkpoint-1630/model-00004-of-00004.safetensors
 create mode 100644 checkpoint-1630/model.safetensors.index.json
 create mode 100644 checkpoint-1630/rng_state_0.pth
 create mode 100644 checkpoint-1630/rng_state_1.pth
 create mode 100644 checkpoint-1630/rng_state_10.pth
 create mode 100644 checkpoint-1630/rng_state_11.pth
 create mode 100644 checkpoint-1630/rng_state_12.pth
 create mode 100644 checkpoint-1630/rng_state_13.pth
 create mode 100644 checkpoint-1630/rng_state_14.pth
 create mode 100644 checkpoint-1630/rng_state_15.pth
 create mode 100644 checkpoint-1630/rng_state_16.pth
 create mode 100644 checkpoint-1630/rng_state_17.pth
 create mode 100644 checkpoint-1630/rng_state_18.pth
 create mode 100644 checkpoint-1630/rng_state_19.pth
 create mode 100644 checkpoint-1630/rng_state_2.pth
 create mode 100644 checkpoint-1630/rng_state_20.pth
 create mode 100644 checkpoint-1630/rng_state_21.pth
 create mode 100644 checkpoint-1630/rng_state_22.pth
 create mode 100644 checkpoint-1630/rng_state_23.pth
 create mode 100644 checkpoint-1630/rng_state_24.pth
 create mode 100644 checkpoint-1630/rng_state_25.pth
 create mode 100644 checkpoint-1630/rng_state_26.pth
 create mode 100644 checkpoint-1630/rng_state_27.pth
 create mode 100644 checkpoint-1630/rng_state_28.pth
 create mode 100644 checkpoint-1630/rng_state_29.pth
 create mode 100644 checkpoint-1630/rng_state_3.pth
 create mode 100644 checkpoint-1630/rng_state_30.pth
 create mode 100644 checkpoint-1630/rng_state_31.pth
 create mode 100644 checkpoint-1630/rng_state_4.pth
 create mode 100644 checkpoint-1630/rng_state_5.pth
 create mode 100644 checkpoint-1630/rng_state_6.pth
 create mode 100644 checkpoint-1630/rng_state_7.pth
 create mode 100644 checkpoint-1630/rng_state_8.pth
 create mode 100644 checkpoint-1630/rng_state_9.pth
 create mode 100644 checkpoint-1630/scheduler.pt
 create mode 100644 checkpoint-1630/special_tokens_map.json
 create mode 100644 checkpoint-1630/tokenizer.json
 create mode 100644 checkpoint-1630/tokenizer_config.json
 create mode 100644 checkpoint-1630/trainer_state.json
 create mode 100644 checkpoint-1630/training_args.bin
 create mode 100644 checkpoint-1630/zero_to_fp32.py
 create mode 100644 checkpoint-2445/config.json
 create mode 100644 checkpoint-2445/generation_config.json
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_0_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_10_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_11_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_12_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_13_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_14_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_15_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_16_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_17_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_18_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_19_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_1_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_20_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_21_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_22_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_23_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_24_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_25_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_26_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_27_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_28_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_29_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_2_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_30_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_31_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_3_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_4_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_5_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_6_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_7_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_8_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/global_step2445/zero_pp_rank_9_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-2445/latest
 create mode 100644 checkpoint-2445/model-00001-of-00004.safetensors
 create mode 100644 checkpoint-2445/model-00002-of-00004.safetensors
 create mode 100644 checkpoint-2445/model-00003-of-00004.safetensors
 create mode 100644 checkpoint-2445/model-00004-of-00004.safetensors
 create mode 100644 checkpoint-2445/model.safetensors.index.json
 create mode 100644 checkpoint-2445/rng_state_0.pth
 create mode 100644 checkpoint-2445/rng_state_1.pth
 create mode 100644 checkpoint-2445/rng_state_10.pth
 create mode 100644 checkpoint-2445/rng_state_11.pth
 create mode 100644 checkpoint-2445/rng_state_12.pth
 create mode 100644 checkpoint-2445/rng_state_13.pth
 create mode 100644 checkpoint-2445/rng_state_14.pth
 create mode 100644 checkpoint-2445/rng_state_15.pth
 create mode 100644 checkpoint-2445/rng_state_16.pth
 create mode 100644 checkpoint-2445/rng_state_17.pth
 create mode 100644 checkpoint-2445/rng_state_18.pth
 create mode 100644 checkpoint-2445/rng_state_19.pth
 create mode 100644 checkpoint-2445/rng_state_2.pth
 create mode 100644 checkpoint-2445/rng_state_20.pth
 create mode 100644 checkpoint-2445/rng_state_21.pth
 create mode 100644 checkpoint-2445/rng_state_22.pth
 create mode 100644 checkpoint-2445/rng_state_23.pth
 create mode 100644 checkpoint-2445/rng_state_24.pth
 create mode 100644 checkpoint-2445/rng_state_25.pth
 create mode 100644 checkpoint-2445/rng_state_26.pth
 create mode 100644 checkpoint-2445/rng_state_27.pth
 create mode 100644 checkpoint-2445/rng_state_28.pth
 create mode 100644 checkpoint-2445/rng_state_29.pth
 create mode 100644 checkpoint-2445/rng_state_3.pth
 create mode 100644 checkpoint-2445/rng_state_30.pth
 create mode 100644 checkpoint-2445/rng_state_31.pth
 create mode 100644 checkpoint-2445/rng_state_4.pth
 create mode 100644 checkpoint-2445/rng_state_5.pth
 create mode 100644 checkpoint-2445/rng_state_6.pth
 create mode 100644 checkpoint-2445/rng_state_7.pth
 create mode 100644 checkpoint-2445/rng_state_8.pth
 create mode 100644 checkpoint-2445/rng_state_9.pth
 create mode 100644 checkpoint-2445/scheduler.pt
 create mode 100644 checkpoint-2445/special_tokens_map.json
 create mode 100644 checkpoint-2445/tokenizer.json
 create mode 100644 checkpoint-2445/tokenizer_config.json
 create mode 100644 checkpoint-2445/trainer_state.json
 create mode 100644 checkpoint-2445/training_args.bin
 create mode 100644 checkpoint-2445/zero_to_fp32.py
 create mode 100644 checkpoint-815/config.json
 create mode 100644 checkpoint-815/generation_config.json
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_0_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_10_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_11_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_12_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_13_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_14_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_15_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_16_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_17_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_18_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_19_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_1_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_20_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_21_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_22_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_23_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_24_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_25_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_26_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_27_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_28_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_29_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_2_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_30_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_31_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_3_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_4_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_5_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_6_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_7_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_8_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/global_step815/zero_pp_rank_9_mp_rank_00_model_states.pt
 create mode 100644 checkpoint-815/latest
 create mode 100644 checkpoint-815/model-00001-of-00004.safetensors
 create mode 100644 checkpoint-815/model-00002-of-00004.safetensors
 create mode 100644 checkpoint-815/model-00003-of-00004.safetensors
 create mode 100644 checkpoint-815/model-00004-of-00004.safetensors
 create mode 100644 checkpoint-815/model.safetensors.index.json
 create mode 100644 checkpoint-815/rng_state_0.pth
 create mode 100644 checkpoint-815/rng_state_1.pth
 create mode 100644 checkpoint-815/rng_state_10.pth
 create mode 100644 checkpoint-815/rng_state_11.pth
 create mode 100644 checkpoint-815/rng_state_12.pth
 create mode 100644 checkpoint-815/rng_state_13.pth
 create mode 100644 checkpoint-815/rng_state_14.pth
 create mode 100644 checkpoint-815/rng_state_15.pth
 create mode 100644 checkpoint-815/rng_state_16.pth
 create mode 100644 checkpoint-815/rng_state_17.pth
 create mode 100644 checkpoint-815/rng_state_18.pth
 create mode 100644 checkpoint-815/rng_state_19.pth
 create mode 100644 checkpoint-815/rng_state_2.pth
 create mode 100644 checkpoint-815/rng_state_20.pth
 create mode 100644 checkpoint-815/rng_state_21.pth
 create mode 100644 checkpoint-815/rng_state_22.pth
 create mode 100644 checkpoint-815/rng_state_23.pth
 create mode 100644 checkpoint-815/rng_state_24.pth
 create mode 100644 checkpoint-815/rng_state_25.pth
 create mode 100644 checkpoint-815/rng_state_26.pth
 create mode 100644 checkpoint-815/rng_state_27.pth
 create mode 100644 checkpoint-815/rng_state_28.pth
 create mode 100644 checkpoint-815/rng_state_29.pth
 create mode 100644 checkpoint-815/rng_state_3.pth
 create mode 100644 checkpoint-815/rng_state_30.pth
 create mode 100644 checkpoint-815/rng_state_31.pth
 create mode 100644 checkpoint-815/rng_state_4.pth
 create mode 100644 checkpoint-815/rng_state_5.pth
 create mode 100644 checkpoint-815/rng_state_6.pth
 create mode 100644 checkpoint-815/rng_state_7.pth
 create mode 100644 checkpoint-815/rng_state_8.pth
 create mode 100644 checkpoint-815/rng_state_9.pth
 create mode 100644 checkpoint-815/scheduler.pt
 create mode 100644 checkpoint-815/special_tokens_map.json
 create mode 100644 checkpoint-815/tokenizer.json
 create mode 100644 checkpoint-815/tokenizer_config.json
 create mode 100644 checkpoint-815/trainer_state.json
 create mode 100644 checkpoint-815/training_args.bin
 create mode 100644 checkpoint-815/zero_to_fp32.py
 create mode 100644 config.json
 create mode 100644 configuration.json
 create mode 100644 generation_config.json
 create mode 100644 model-00001-of-00004.safetensors
 create mode 100644 model-00002-of-00004.safetensors
 create mode 100644 model-00003-of-00004.safetensors
 create mode 100644 model-00004-of-00004.safetensors
 create mode 100644 model.safetensors.index.json
 create mode 100644 runs/Apr03_06-04-06_ip-10-1-21-143/events.out.tfevents.1743660284.ip-10-1-21-143.961514.0
 create mode 100644 special_tokens_map.json
 create mode 100644 tokenizer.json
 create mode 100644 tokenizer_config.json
 create mode 100644 training_args.bin

diff --git a/.gitattributes b/.gitattributes
new file mode 100644
index 0000000..23c1e42
--- /dev/null
+++ b/.gitattributes
@@ -0,0 +1,52 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bin.* filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zstandard filter=lfs diff=lfs merge=lfs -text
+*.tfevents* filter=lfs diff=lfs merge=lfs -text
+*.db* filter=lfs diff=lfs merge=lfs -text
+*.ark* filter=lfs diff=lfs merge=lfs -text
+**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text
+**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text
+**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.gguf* filter=lfs diff=lfs merge=lfs -text
+*.ggml filter=lfs diff=lfs merge=lfs -text
+*.llamafile* filter=lfs diff=lfs merge=lfs -text
+*.pt2 filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+
+checkpoint-1630/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-2445/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-815/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text
\ No newline at end of file
diff --git a/README.md b/README.md
new file mode 100644
index 0000000..7915238
--- /dev/null
+++ b/README.md
@@ -0,0 +1,66 @@
+---
+library_name: transformers
+model_name: outputs/ablation-145-a128.dpo.armorm.rp.tl.8e7-shisa-v2-llama-3.1-8b
+tags:
+- generated_from_trainer
+licence: license
+---
+
+# Model Card for outputs/ablation-145-a128.dpo.armorm.rp.tl.8e7-shisa-v2-llama-3.1-8b
+
+This model is a fine-tuned version of [None](https://huggingface.co/None).
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="None", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/augmxnt/shisa-v2/runs/hximnojs) 
+
+
+This model was trained with DPO, a method introduced in [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://huggingface.co/papers/2305.18290).
+
+### Framework versions
+
+- TRL: 0.15.1
+- Transformers: 4.50.0
+- Pytorch: 2.6.0
+- Datasets: 3.4.1
+- Tokenizers: 0.21.1
+
+## Citations
+
+Cite DPO as:
+
+```bibtex
+@inproceedings{rafailov2023direct,
+    title        = {{Direct Preference Optimization: Your Language Model is Secretly a Reward Model}},
+    author       = {Rafael Rafailov and Archit Sharma and Eric Mitchell and Christopher D. Manning and Stefano Ermon and Chelsea Finn},
+    year         = 2023,
+    booktitle    = {Advances in Neural Information Processing Systems 36: Annual Conference on Neural Information Processing Systems 2023, NeurIPS 2023, New Orleans, LA, USA, December 10 - 16, 2023},
+    url          = {http://papers.nips.cc/paper_files/paper/2023/hash/a85b405ed65c6477a4fe8302b5e06ce7-Abstract-Conference.html},
+    editor       = {Alice Oh and Tristan Naumann and Amir Globerson and Kate Saenko and Moritz Hardt and Sergey Levine},
+}
+```
+
+Cite TRL as:
+    
+```bibtex
+@misc{vonwerra2022trl,
+	title        = {{TRL: Transformer Reinforcement Learning}},
+	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
+	year         = 2020,
+	journal      = {GitHub repository},
+	publisher    = {GitHub},
+	howpublished = {\url{https://github.com/huggingface/trl}}
+}
+```
\ No newline at end of file
diff --git a/checkpoint-1630/config.json b/checkpoint-1630/config.json
new file mode 100644
index 0000000..ded5a8b
--- /dev/null
+++ b/checkpoint-1630/config.json
@@ -0,0 +1,35 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128009,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 8.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.50.0",
+  "use_cache": false,
+  "vocab_size": 128256
+}
diff --git a/checkpoint-1630/generation_config.json b/checkpoint-1630/generation_config.json
new file mode 100644
index 0000000..0acdb5a
--- /dev/null
+++ b/checkpoint-1630/generation_config.json
@@ -0,0 +1,12 @@
+{
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.50.0"
+}
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..d3136f7
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f2ff751ecb4d9b0c48c9f3ab30c850a3e47d86cbbf06ad7e07a7667aeacc229
+size 1514015111
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..0e3b4ca
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62568e0ae2360c990c393491cb7e0638a4aa679b0a4b726c62bb23e140cef788
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..82133df
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ab2c6f52fbfab51728e350211c9d8a2c50ad1fab85a29af40f737f6ef8a7486
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..f50cc9f
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3228b0f9e68b07feacf1ec74b58a77734a6833a56feff58a5fa5ae6a322bdb8e
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..b9255df
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:182f66ca2925a08d69d3d715f5a2050c5f585efb4e674b18d3de7ef20d7b82bb
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..f44a53b
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbca7be0ab555b846f4f52e456dc6b7087e9e873bcb4e6060a0c49efdf91989f
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..0090fa8
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8951ea2efd81817fdd454eda1b889545038c57cc818a366ec277f34c3629f536
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..b2295bc
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9d1f28f8dfd0852887c0a8152f0779e862b1c75972e91f52f8c477d01b98a10
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..c3aa89e
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f8bdc7351c4ec70c735758942c217629a531f9ba50026881249a8ec1f8b3ae0
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..240cd23
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a26066feea8040450b7bd4123a3e0387e590d653add2ac4f4db08f6effdc255a
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..bf835dd
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bae66ce741625baa8bd8c628fd93f931537d226d62df49f7cc57c645309a9ad1
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..89b8ba9
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd8311e8dbafc2442ea841b9de9e4e909414dc4c309b2072e07da036a9cf11e5
+size 1514015111
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..ec5a068
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:350d3f8ea39c9dab165fe599abc9167a9dbdaa6a7675b17fd589953b677bcfc9
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..79799a4
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c6e4a01a1c89bf524dac6a8462d7a8a083a1e21665933cd211e21f969ef641f
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..fb6d771
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23baed0ac03f5f61f4345076593f2c099b61cb0c6dadc3eccaf3b88193d347cb
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..29f1c58
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96d47f3e3ade9770cb742ba669392589fef7e589c30cb7070038dc1269fdff34
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..f49793d
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fa1245a951da78af6cacb04072ff6f57b2304ac19a2b89464bff73779d057e5
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..896a044
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25b7852718d4f8b104fc3789194035893a209e5dc0ca0d94110e8ac5b74cc59f
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..a6ba639
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d32426d0df62bcb67fc700f034c32fd74df1c8847e8a2b92814ee399b77a02e8
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..0cf3db9
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f68765512d4c6f36dd48ed36c9adc8665bede13ef954ed138d8a4fbf1b894fb
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..f805449
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:348915fcf751af804643febb426f11055c44b0791f31a81c193c15428f06f4bd
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..7e52c69
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e3adda7eaffe1b9c378a19e62fd2664e2778e7f061a9a86cc64c7a1a685dc88
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..ec82f51
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c50fb69a632718fe13f5be2c4a0de38fc8320e17d7ed2f712b967959d570e59b
+size 1514015111
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..423936a
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c0f5273defe1299613121c638bc9015b5f54e2063dfc398fa76e7139763c4d5
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..d3e1883
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22335c0a64d49089b34c66b9c5531abc9a0ea09b3b8ca758f75a98040954896d
+size 1514016442
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..b2d324b
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23e7520ff9578f147ba068f92f83dbb7115fd01f47c675d45224e930fa495d6e
+size 1514015111
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..8b69274
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6498a0637ee46668786e91053528e9d5d0ef24bb8f93a91267ee4ccfd5bd11d8
+size 1514015111
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..7b3913c
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e0108920635ba4089432970e8956e05e2a30a37ac2c0969288108ad5428d04c3
+size 1514015111
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..d96cd62
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54dc737e2af610bf711924774adcfbeae4bff8131093b054a7c6f14d179941a5
+size 1514015111
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..8ca3a4a
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:657f0458e96606e47586b5d7193728fec329c24d3ec50c9d8025184dac16c993
+size 1514015111
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..afdae2e
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1cc2bba5c8673d60b336e1ec8c0505693d7f00efff77fd35cbde5511bf07dc2e
+size 1514015111
diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..588370d
--- /dev/null
+++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c663f1fc963b83fb59eabc1f2c2bfb0bc5bf2ce50b72440957ea3ef241878572
+size 1514015111
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_0_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_0_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..db997e4
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_0_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa532a7c8189bc930aebd05eba451a267458553a6f3af3bb3b1e98d8523d76be
+size 150245
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_10_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_10_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..aa9c61b
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_10_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f515b12cf75954828dd92c701dcb04425ca7da970eaf6a88653baf1d62011269
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_11_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_11_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..7d5ca98
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_11_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84c969c26f5eecbc4c2484d80c3ea790de2bb535e0d02d9935d79c357fbb8f86
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_12_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_12_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..c8b5fd8
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_12_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bafcd493f85462e19b9224677c7514bb3b94f423861f053cc4c3fa2be59a53f8
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_13_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_13_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..e5d4fff
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_13_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c8717e3c8eaee90ce1e24d39bd0d4d19bc5423035c1a8041bb31c3c2f460158
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_14_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_14_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..9740439
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_14_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94de6015b071a4de07754007a639c92a62d94046a025359225e936dcb362950c
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_15_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_15_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..76bfa5b
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_15_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc643726424c61753779b64ec836f54d53bc3e54c14a897e082cbd4cc1ac39cc
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_16_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_16_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..3fe6b8d
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_16_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45d822837deac3b5b69558d5eb8bf16f733c248644913ba5d1387dfcdbb246e2
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_17_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_17_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..141ff82
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_17_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d83791ff8b0cc02d65fd9e6c0250fee80c813bfa93bc74440d0d9546b742e731
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_18_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_18_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..67fbf8e
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_18_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1ad648ff2f6632d194310cd590612a34dcee5ae59afc2705330de48f48818fa
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_19_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_19_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..c8ebd46
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_19_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5afb79ef6850bc18a5d736840c68a776e930087b38321ffc62455c3ca324957
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_1_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_1_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..45d83f1
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_1_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:31aa88b36939fbb7eb1353cfb81e6285e052452a3be3f5322a5898606ab6f099
+size 150245
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_20_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_20_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..a46f0f7
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_20_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f78a8d21611157ba84d8e933650e4314450e598ef9390e2ef140ffbaebcbb413
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_21_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_21_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..7784a83
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_21_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbd5d9d4ba33afd5b8681d99f7b557498a6bc34b1cf469cd6fdfad70ef3ba462
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_22_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_22_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..5bf2872
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_22_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce1032af673079a83985ef1e3221449808e2e402a401e80878f19ebdefc1fb91
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_23_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_23_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..458fd3e
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_23_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93c819f2b87b13c7d0365314a8793830cbc3537fce028f88a23171c0a8d139d3
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_24_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_24_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..fbce9f2
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_24_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c15e8195851401da7054ba176dee771b7dace27c96d02c3563ed4787041a24c0
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_25_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_25_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..ecf2083
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_25_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4812cff1c52c194835de7c60b4db6ed5af2c99a3eebf7e1e36a50487b485500
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_26_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_26_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..89980fc
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_26_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1bbdcfe71f050b662d5f093662a16b6fbc07cb1581a11a143e442e34d69d6773
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_27_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_27_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..4fd1656
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_27_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fc965279bf3f1fc656157fe0f3a6ae7d7eff93f4ae7191e81aeb950509b5028
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_28_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_28_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..6fbaaa7
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_28_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11fccd0874abe40f5a9c85f018dd6e8fe275cf1ce335e18aa18d328dfcdafe6c
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_29_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_29_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..d616878
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_29_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d078fc60c854c4f513f1bae8c532e41f3ee95fa8a1627dc892bb71f7eed7b07e
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_2_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_2_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..bdd2801
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_2_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c7765bdf55f1a5a1c6e3f10b1fab6afd8bfe9882aa502f584ecb39326329cd5
+size 150245
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_30_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_30_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..86eb18a
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_30_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd4fc4625ba7ba5d268581dcd1621b8fad50e98dc1cb0a51ee2d310cf38180f9
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_31_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_31_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..2550b22
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_31_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e9dac2bfd65cf28544ef2c368df6588cb38a0e2aea7ca7309e34803e309a869
+size 150540
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_3_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_3_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..5ddc7c4
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_3_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cecb7ce13aae08d3535fed9bc97e11d0a139c2209edf77f3efb1dce2a347b76f
+size 150245
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_4_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_4_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..d4c6a9b
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_4_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4d1df2d1673f71716cd32eac2fdfeeb293b87695c982654cd52881de39b8609
+size 150245
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_5_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_5_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..a04e871
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_5_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4097f82f363f57deca37fc2e94f9db9595a0d3fbff7471e6db2e4996247ea34e
+size 150245
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_6_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_6_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..0a82596
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_6_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8804c594d8aa48eae8a689acf3aa847a2780956548a12706e750c60449febabd
+size 150245
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_7_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_7_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..98edf43
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_7_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9be01bda51e54d84107f28aad68934ece7bc9243d22f37bca36d70d5a687c306
+size 150245
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_8_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_8_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..44dc7c4
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_8_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:965dc1bd2b8bc46f0492c744ea8d80fe74acc33c0a82031d484658924a9188c5
+size 150245
diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_9_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_9_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..aba92fd
--- /dev/null
+++ b/checkpoint-1630/global_step1630/zero_pp_rank_9_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:98208579a411498d49a4075e894283e5b951eee51cec29ca8f3c6dd17dac132c
+size 150245
diff --git a/checkpoint-1630/latest b/checkpoint-1630/latest
new file mode 100644
index 0000000..b876efb
--- /dev/null
+++ b/checkpoint-1630/latest
@@ -0,0 +1 @@
+global_step1630
\ No newline at end of file
diff --git a/checkpoint-1630/model-00001-of-00004.safetensors b/checkpoint-1630/model-00001-of-00004.safetensors
new file mode 100644
index 0000000..635d669
--- /dev/null
+++ b/checkpoint-1630/model-00001-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44ff7867de3d675221920938964b5026d577b4f5f02a02214e13270da75bcd30
+size 4976698672
diff --git a/checkpoint-1630/model-00002-of-00004.safetensors b/checkpoint-1630/model-00002-of-00004.safetensors
new file mode 100644
index 0000000..8b69277
--- /dev/null
+++ b/checkpoint-1630/model-00002-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6d8b74dca61ecc61358c1f5b44695c186de717af5892887a5d5254a7783e5ff7
+size 4999802720
diff --git a/checkpoint-1630/model-00003-of-00004.safetensors b/checkpoint-1630/model-00003-of-00004.safetensors
new file mode 100644
index 0000000..ec09657
--- /dev/null
+++ b/checkpoint-1630/model-00003-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00ff6497ca58336b8344c3f7533d28c4d0b0fe9ca3e709374d135074b38de37b
+size 4915916176
diff --git a/checkpoint-1630/model-00004-of-00004.safetensors b/checkpoint-1630/model-00004-of-00004.safetensors
new file mode 100644
index 0000000..9974cd0
--- /dev/null
+++ b/checkpoint-1630/model-00004-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:31bc897114b9debeae876b5f326483cef021b6a96112628cd759f7275062b21c
+size 1168138808
diff --git a/checkpoint-1630/model.safetensors.index.json b/checkpoint-1630/model.safetensors.index.json
new file mode 100644
index 0000000..0fd8120
--- /dev/null
+++ b/checkpoint-1630/model.safetensors.index.json
@@ -0,0 +1,298 @@
+{
+  "metadata": {
+    "total_size": 16060522496
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.norm.weight": "model-00004-of-00004.safetensors"
+  }
+}
diff --git a/checkpoint-1630/rng_state_0.pth b/checkpoint-1630/rng_state_0.pth
new file mode 100644
index 0000000..b346349
--- /dev/null
+++ b/checkpoint-1630/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad8a35afd8967cbb748405387e44426e43ad127028e826eddc9b67d2ca873c85
+size 15984
diff --git a/checkpoint-1630/rng_state_1.pth b/checkpoint-1630/rng_state_1.pth
new file mode 100644
index 0000000..68f3c69
--- /dev/null
+++ b/checkpoint-1630/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f338ce80d7c441076bfc8c53b84067a0181f5a14e80c13d5acb8150b659f4d73
+size 15984
diff --git a/checkpoint-1630/rng_state_10.pth b/checkpoint-1630/rng_state_10.pth
new file mode 100644
index 0000000..09612af
--- /dev/null
+++ b/checkpoint-1630/rng_state_10.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d4c67c1d0ec6b889370f634ff29f584f829acaeaf69196a69304fdba936f9d7
+size 15997
diff --git a/checkpoint-1630/rng_state_11.pth b/checkpoint-1630/rng_state_11.pth
new file mode 100644
index 0000000..06a46cd
--- /dev/null
+++ b/checkpoint-1630/rng_state_11.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afd0c4e2ac5f61d5f4df5d9d5783018b7937173b4afdbb92816c2668982cb351
+size 15997
diff --git a/checkpoint-1630/rng_state_12.pth b/checkpoint-1630/rng_state_12.pth
new file mode 100644
index 0000000..bf6f992
--- /dev/null
+++ b/checkpoint-1630/rng_state_12.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9d1c7f561c5512193e8053ecf371e4fef647a8503481f2dad76332ddbe164fd
+size 15997
diff --git a/checkpoint-1630/rng_state_13.pth b/checkpoint-1630/rng_state_13.pth
new file mode 100644
index 0000000..eac9c7f
--- /dev/null
+++ b/checkpoint-1630/rng_state_13.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7604a3cfaf0e3c3f2aa2b3547660615db82205d4158951bb90f69ffc560ed179
+size 15997
diff --git a/checkpoint-1630/rng_state_14.pth b/checkpoint-1630/rng_state_14.pth
new file mode 100644
index 0000000..29fb251
--- /dev/null
+++ b/checkpoint-1630/rng_state_14.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d984186a3bc06445d6c077fcbf64da96d47d3ed704e68f0212045efb4b91dbb
+size 15997
diff --git a/checkpoint-1630/rng_state_15.pth b/checkpoint-1630/rng_state_15.pth
new file mode 100644
index 0000000..0bf361b
--- /dev/null
+++ b/checkpoint-1630/rng_state_15.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b3945ba3535361fa13a8460a70ebbd44ece2514a1c9dcb76abfe6f54a775c60
+size 15997
diff --git a/checkpoint-1630/rng_state_16.pth b/checkpoint-1630/rng_state_16.pth
new file mode 100644
index 0000000..0e7adea
--- /dev/null
+++ b/checkpoint-1630/rng_state_16.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:853a02543fd3223edd9334ea75c62e6e99ded22ac864e9930fbf5add26045fe7
+size 15997
diff --git a/checkpoint-1630/rng_state_17.pth b/checkpoint-1630/rng_state_17.pth
new file mode 100644
index 0000000..18e8350
--- /dev/null
+++ b/checkpoint-1630/rng_state_17.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c0420ad6cf62c9fe9cda1807f68e8e40235a8917dbab858f800d24e3ba8eeae
+size 15997
diff --git a/checkpoint-1630/rng_state_18.pth b/checkpoint-1630/rng_state_18.pth
new file mode 100644
index 0000000..9b6c223
--- /dev/null
+++ b/checkpoint-1630/rng_state_18.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e167524b369e20a3a17ee7f6124322da680ccf2ae553b4adc0e7462d2dfba82b
+size 15997
diff --git a/checkpoint-1630/rng_state_19.pth b/checkpoint-1630/rng_state_19.pth
new file mode 100644
index 0000000..b0d318f
--- /dev/null
+++ b/checkpoint-1630/rng_state_19.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0915a78eba7c1d2768ecd03818e3c01afd7f3168a3318b1212413bd2c62f7f1
+size 15997
diff --git a/checkpoint-1630/rng_state_2.pth b/checkpoint-1630/rng_state_2.pth
new file mode 100644
index 0000000..be044f6
--- /dev/null
+++ b/checkpoint-1630/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9fbc9fa428939be10b46779f0eb5cd833e0da426b1cbdee77b3a55b6952235b
+size 15984
diff --git a/checkpoint-1630/rng_state_20.pth b/checkpoint-1630/rng_state_20.pth
new file mode 100644
index 0000000..ece204f
--- /dev/null
+++ b/checkpoint-1630/rng_state_20.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f8f2fdb90584ebb7247c85570082ac7c6338216de82b07f132a2c679923379a
+size 15997
diff --git a/checkpoint-1630/rng_state_21.pth b/checkpoint-1630/rng_state_21.pth
new file mode 100644
index 0000000..5b62c48
--- /dev/null
+++ b/checkpoint-1630/rng_state_21.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a3863ac6512e946a826d479520004385d29f3fd0a2b686ccbfdf488332aa442
+size 15997
diff --git a/checkpoint-1630/rng_state_22.pth b/checkpoint-1630/rng_state_22.pth
new file mode 100644
index 0000000..b0fa34f
--- /dev/null
+++ b/checkpoint-1630/rng_state_22.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1dfe94fb47194a2a298bed1b33ad18c2132747148814886bd9b8910a7691ba0b
+size 15997
diff --git a/checkpoint-1630/rng_state_23.pth b/checkpoint-1630/rng_state_23.pth
new file mode 100644
index 0000000..c7c7ed4
--- /dev/null
+++ b/checkpoint-1630/rng_state_23.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c683c09c8883037948fabb143cd972207db3cee22d7efcbcc5866a8ec85c94c6
+size 15997
diff --git a/checkpoint-1630/rng_state_24.pth b/checkpoint-1630/rng_state_24.pth
new file mode 100644
index 0000000..be36a1a
--- /dev/null
+++ b/checkpoint-1630/rng_state_24.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:53c201c89debf1ca694786330ea98ca69da6340ac3b70f40aad77213d1771bf8
+size 15997
diff --git a/checkpoint-1630/rng_state_25.pth b/checkpoint-1630/rng_state_25.pth
new file mode 100644
index 0000000..b34b54b
--- /dev/null
+++ b/checkpoint-1630/rng_state_25.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b7b8393dc10fb6ab2222179b268184fa5330fd417b66ebb65986818f47334a55
+size 15997
diff --git a/checkpoint-1630/rng_state_26.pth b/checkpoint-1630/rng_state_26.pth
new file mode 100644
index 0000000..354820a
--- /dev/null
+++ b/checkpoint-1630/rng_state_26.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b29094f6dcc92be7d58da39df551c573728836c5aec67bef90790810c0d3882
+size 15997
diff --git a/checkpoint-1630/rng_state_27.pth b/checkpoint-1630/rng_state_27.pth
new file mode 100644
index 0000000..747b806
--- /dev/null
+++ b/checkpoint-1630/rng_state_27.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b64ff98dfb30bbfa08760a0a79d9be7407917ff0e32f22b03d6528bf7e292e44
+size 15997
diff --git a/checkpoint-1630/rng_state_28.pth b/checkpoint-1630/rng_state_28.pth
new file mode 100644
index 0000000..b316dea
--- /dev/null
+++ b/checkpoint-1630/rng_state_28.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04e19d4d07b899c9158590672b9acc125c3fb4082f5aed08ddaee573b4dbf33f
+size 15997
diff --git a/checkpoint-1630/rng_state_29.pth b/checkpoint-1630/rng_state_29.pth
new file mode 100644
index 0000000..7b0fc3c
--- /dev/null
+++ b/checkpoint-1630/rng_state_29.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47a620c2380d984b51262d8f8d1853631c1bd13c3da4ea8f73e8dc61dd3fb765
+size 15997
diff --git a/checkpoint-1630/rng_state_3.pth b/checkpoint-1630/rng_state_3.pth
new file mode 100644
index 0000000..fc82524
--- /dev/null
+++ b/checkpoint-1630/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac55dba0b79d5fa4699d239da2f966d52040d576d31234ac8d4632e6956481bc
+size 15984
diff --git a/checkpoint-1630/rng_state_30.pth b/checkpoint-1630/rng_state_30.pth
new file mode 100644
index 0000000..f88acd7
--- /dev/null
+++ b/checkpoint-1630/rng_state_30.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1516e58b9c7537b0c41d41487bc1464770e53c3dabdb8cf4f5c07bcc51e07ed6
+size 15997
diff --git a/checkpoint-1630/rng_state_31.pth b/checkpoint-1630/rng_state_31.pth
new file mode 100644
index 0000000..4e033f4
--- /dev/null
+++ b/checkpoint-1630/rng_state_31.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58a50bef57066132c2e5b6b53fbb78e101505dd4ad585ec52fb3c11ca0663c22
+size 15997
diff --git a/checkpoint-1630/rng_state_4.pth b/checkpoint-1630/rng_state_4.pth
new file mode 100644
index 0000000..d30f52a
--- /dev/null
+++ b/checkpoint-1630/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af2d0c015100768ffa23faf3b6c2d54ea89eb045603e30e55cd211e06ff34972
+size 15984
diff --git a/checkpoint-1630/rng_state_5.pth b/checkpoint-1630/rng_state_5.pth
new file mode 100644
index 0000000..c8715d2
--- /dev/null
+++ b/checkpoint-1630/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c60a1b40608e34bc801c8231f97b81c53b5290dfaed1b9cd0ccbeca29574a991
+size 15984
diff --git a/checkpoint-1630/rng_state_6.pth b/checkpoint-1630/rng_state_6.pth
new file mode 100644
index 0000000..1ed791b
--- /dev/null
+++ b/checkpoint-1630/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ad6a142a403eb9aafc4a3a9a856bca648fe31fd22d796867baca31fb13656aa
+size 15984
diff --git a/checkpoint-1630/rng_state_7.pth b/checkpoint-1630/rng_state_7.pth
new file mode 100644
index 0000000..800c3bb
--- /dev/null
+++ b/checkpoint-1630/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38bc23a138cc800b22881742c0f3f9a71731a9a7111c6058a0077e6274d21773
+size 15984
diff --git a/checkpoint-1630/rng_state_8.pth b/checkpoint-1630/rng_state_8.pth
new file mode 100644
index 0000000..d7068b4
--- /dev/null
+++ b/checkpoint-1630/rng_state_8.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4b75777b43df70dca87f91deed2e613941188792f244d5eeb06da8ab038bd36
+size 15984
diff --git a/checkpoint-1630/rng_state_9.pth b/checkpoint-1630/rng_state_9.pth
new file mode 100644
index 0000000..d42657d
--- /dev/null
+++ b/checkpoint-1630/rng_state_9.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8f9eb93d0ac349ba8b2271c417e5fcc4b70447a424f3f861c3338fbd99a48f4
+size 15984
diff --git a/checkpoint-1630/scheduler.pt b/checkpoint-1630/scheduler.pt
new file mode 100644
index 0000000..8a87b10
--- /dev/null
+++ b/checkpoint-1630/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:12f582ff1a2b433c6b1629abd4aa84fd7ced4be10cd95fc7aa54d27450a3470f
+size 1064
diff --git a/checkpoint-1630/special_tokens_map.json b/checkpoint-1630/special_tokens_map.json
new file mode 100644
index 0000000..278b7f0
--- /dev/null
+++ b/checkpoint-1630/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-1630/tokenizer.json b/checkpoint-1630/tokenizer.json
new file mode 100644
index 0000000..1c1d8d5
--- /dev/null
+++ b/checkpoint-1630/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-1630/tokenizer_config.json b/checkpoint-1630/tokenizer_config.json
new file mode 100644
index 0000000..8098161
--- /dev/null
+++ b/checkpoint-1630/tokenizer_config.json
@@ -0,0 +1,2064 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "<|end_of_text|>",
+  "tokenizer_class": "PreTrainedTokenizer"
+}
diff --git a/checkpoint-1630/trainer_state.json b/checkpoint-1630/trainer_state.json
new file mode 100644
index 0000000..1d31dfe
--- /dev/null
+++ b/checkpoint-1630/trainer_state.json
@@ -0,0 +1,24484 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 1630,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001226993865030675,
+      "grad_norm": 151.42932274482283,
+      "learning_rate": 3.2e-08,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -474.0,
+      "logps/rejected": -316.0,
+      "loss": 0.6914,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.00245398773006135,
+      "grad_norm": 162.73344539002164,
+      "learning_rate": 6.4e-08,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -292.0,
+      "loss": 0.6914,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 2
+    },
+    {
+      "epoch": 0.0036809815950920245,
+      "grad_norm": 111.8042794987884,
+      "learning_rate": 9.599999999999999e-08,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -272.0,
+      "loss": 0.6886,
+      "rewards/accuracies": 0.3046875,
+      "rewards/chosen": 0.01531982421875,
+      "rewards/margins": 0.01361083984375,
+      "rewards/rejected": 0.0016632080078125,
+      "step": 3
+    },
+    {
+      "epoch": 0.0049079754601227,
+      "grad_norm": 155.41451533543852,
+      "learning_rate": 1.28e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -242.0,
+      "loss": 0.7131,
+      "rewards/accuracies": 0.328125,
+      "rewards/chosen": -0.00408935546875,
+      "rewards/margins": -0.0218505859375,
+      "rewards/rejected": 0.017822265625,
+      "step": 4
+    },
+    {
+      "epoch": 0.006134969325153374,
+      "grad_norm": 307.019899938903,
+      "learning_rate": 1.6e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -270.0,
+      "loss": 0.6906,
+      "rewards/accuracies": 0.3515625,
+      "rewards/chosen": 0.01171875,
+      "rewards/margins": 0.01251220703125,
+      "rewards/rejected": -0.00078582763671875,
+      "step": 5
+    },
+    {
+      "epoch": 0.007361963190184049,
+      "grad_norm": 340.66722515261114,
+      "learning_rate": 1.9199999999999997e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -306.0,
+      "loss": 0.7151,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.0228271484375,
+      "rewards/margins": -0.034912109375,
+      "rewards/rejected": 0.01220703125,
+      "step": 6
+    },
+    {
+      "epoch": 0.008588957055214725,
+      "grad_norm": 239.24556126143491,
+      "learning_rate": 2.2400000000000002e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -266.0,
+      "loss": 0.7032,
+      "rewards/accuracies": 0.2734375,
+      "rewards/chosen": -0.0003414154052734375,
+      "rewards/margins": -0.01373291015625,
+      "rewards/rejected": 0.01336669921875,
+      "step": 7
+    },
+    {
+      "epoch": 0.0098159509202454,
+      "grad_norm": 139.49365426176607,
+      "learning_rate": 2.56e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -292.0,
+      "loss": 0.7012,
+      "rewards/accuracies": 0.3671875,
+      "rewards/chosen": 0.0284423828125,
+      "rewards/margins": -0.00927734375,
+      "rewards/rejected": 0.037841796875,
+      "step": 8
+    },
+    {
+      "epoch": 0.011042944785276074,
+      "grad_norm": 122.82829055069632,
+      "learning_rate": 2.88e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -294.0,
+      "loss": 0.6871,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.037353515625,
+      "rewards/margins": 0.0244140625,
+      "rewards/rejected": 0.01300048828125,
+      "step": 9
+    },
+    {
+      "epoch": 0.012269938650306749,
+      "grad_norm": 137.4984961934261,
+      "learning_rate": 3.2e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -278.0,
+      "loss": 0.6605,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": 0.057861328125,
+      "rewards/margins": 0.07275390625,
+      "rewards/rejected": -0.01495361328125,
+      "step": 10
+    },
+    {
+      "epoch": 0.013496932515337423,
+      "grad_norm": 154.04829614927058,
+      "learning_rate": 3.52e-07,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -296.0,
+      "loss": 0.6512,
+      "rewards/accuracies": 0.4765625,
+      "rewards/chosen": 0.0625,
+      "rewards/margins": 0.09912109375,
+      "rewards/rejected": -0.036376953125,
+      "step": 11
+    },
+    {
+      "epoch": 0.014723926380368098,
+      "grad_norm": 292.49015304309825,
+      "learning_rate": 3.8399999999999994e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -262.0,
+      "loss": 0.6321,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": 0.080078125,
+      "rewards/margins": 0.1435546875,
+      "rewards/rejected": -0.06298828125,
+      "step": 12
+    },
+    {
+      "epoch": 0.015950920245398775,
+      "grad_norm": 179.53722097804638,
+      "learning_rate": 4.16e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -249.0,
+      "loss": 0.6194,
+      "rewards/accuracies": 0.6171875,
+      "rewards/chosen": 0.12109375,
+      "rewards/margins": 0.1845703125,
+      "rewards/rejected": -0.0634765625,
+      "step": 13
+    },
+    {
+      "epoch": 0.01717791411042945,
+      "grad_norm": 108.27195029021392,
+      "learning_rate": 4.4800000000000004e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -472.0,
+      "logps/rejected": -312.0,
+      "loss": 0.5477,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": 0.310546875,
+      "rewards/margins": 0.408203125,
+      "rewards/rejected": -0.09814453125,
+      "step": 14
+    },
+    {
+      "epoch": 0.018404907975460124,
+      "grad_norm": 71.36850002174661,
+      "learning_rate": 4.8e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.18359375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -290.0,
+      "loss": 0.5893,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": 0.2412109375,
+      "rewards/margins": 0.337890625,
+      "rewards/rejected": -0.09619140625,
+      "step": 15
+    },
+    {
+      "epoch": 0.0196319018404908,
+      "grad_norm": 80.69275734688873,
+      "learning_rate": 5.12e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -306.0,
+      "loss": 0.5357,
+      "rewards/accuracies": 0.6484375,
+      "rewards/chosen": 0.318359375,
+      "rewards/margins": 0.48828125,
+      "rewards/rejected": -0.1689453125,
+      "step": 16
+    },
+    {
+      "epoch": 0.020858895705521473,
+      "grad_norm": 103.00148200369824,
+      "learning_rate": 5.44e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -255.0,
+      "loss": 0.5279,
+      "rewards/accuracies": 0.6953125,
+      "rewards/chosen": 0.369140625,
+      "rewards/margins": 0.54296875,
+      "rewards/rejected": -0.1728515625,
+      "step": 17
+    },
+    {
+      "epoch": 0.022085889570552148,
+      "grad_norm": 90.51865254706846,
+      "learning_rate": 5.76e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -276.0,
+      "loss": 0.5097,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 0.419921875,
+      "rewards/margins": 0.66015625,
+      "rewards/rejected": -0.23828125,
+      "step": 18
+    },
+    {
+      "epoch": 0.023312883435582823,
+      "grad_norm": 63.171857373704086,
+      "learning_rate": 6.079999999999999e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -290.0,
+      "loss": 0.5237,
+      "rewards/accuracies": 0.6953125,
+      "rewards/chosen": 0.4375,
+      "rewards/margins": 0.80078125,
+      "rewards/rejected": -0.36328125,
+      "step": 19
+    },
+    {
+      "epoch": 0.024539877300613498,
+      "grad_norm": 44.73691218363087,
+      "learning_rate": 6.4e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -462.0,
+      "logps/rejected": -314.0,
+      "loss": 0.4345,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.82421875,
+      "rewards/margins": 1.3984375,
+      "rewards/rejected": -0.5703125,
+      "step": 20
+    },
+    {
+      "epoch": 0.025766871165644172,
+      "grad_norm": 60.19744308670884,
+      "learning_rate": 6.72e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -262.0,
+      "loss": 0.4418,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 1.1328125,
+      "rewards/rejected": -0.50390625,
+      "step": 21
+    },
+    {
+      "epoch": 0.026993865030674847,
+      "grad_norm": 60.06635545400663,
+      "learning_rate": 7.04e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -294.0,
+      "loss": 0.4638,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 0.640625,
+      "rewards/margins": 1.25,
+      "rewards/rejected": -0.61328125,
+      "step": 22
+    },
+    {
+      "epoch": 0.02822085889570552,
+      "grad_norm": 81.66202093067754,
+      "learning_rate": 7.36e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -306.0,
+      "loss": 0.5093,
+      "rewards/accuracies": 0.7421875,
+      "rewards/chosen": 0.81640625,
+      "rewards/margins": 1.421875,
+      "rewards/rejected": -0.6015625,
+      "step": 23
+    },
+    {
+      "epoch": 0.029447852760736196,
+      "grad_norm": 83.23689816485307,
+      "learning_rate": 7.679999999999999e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -278.0,
+      "loss": 0.511,
+      "rewards/accuracies": 0.7265625,
+      "rewards/chosen": 0.47265625,
+      "rewards/margins": 1.03125,
+      "rewards/rejected": -0.5546875,
+      "step": 24
+    },
+    {
+      "epoch": 0.03067484662576687,
+      "grad_norm": 38.90986069898983,
+      "learning_rate": 8e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -434.0,
+      "logps/rejected": -300.0,
+      "loss": 0.4067,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.8515625,
+      "rewards/margins": 1.7578125,
+      "rewards/rejected": -0.90625,
+      "step": 25
+    },
+    {
+      "epoch": 0.03190184049079755,
+      "grad_norm": 47.94846803093583,
+      "learning_rate": 7.996694214876033e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -276.0,
+      "loss": 0.4388,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 1.84375,
+      "rewards/rejected": -0.8203125,
+      "step": 26
+    },
+    {
+      "epoch": 0.033128834355828224,
+      "grad_norm": 44.30203970039667,
+      "learning_rate": 7.993388429752066e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -280.0,
+      "loss": 0.3991,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 2.234375,
+      "rewards/rejected": -1.0,
+      "step": 27
+    },
+    {
+      "epoch": 0.0343558282208589,
+      "grad_norm": 38.710283535206564,
+      "learning_rate": 7.990082644628099e-07,
+      "logits/chosen": -0.06591796875,
+      "logits/rejected": -0.1767578125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -262.0,
+      "loss": 0.4113,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 2.015625,
+      "rewards/rejected": -0.74609375,
+      "step": 28
+    },
+    {
+      "epoch": 0.03558282208588957,
+      "grad_norm": 52.910805586215126,
+      "learning_rate": 7.986776859504132e-07,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -322.0,
+      "loss": 0.392,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 2.546875,
+      "rewards/rejected": -1.1015625,
+      "step": 29
+    },
+    {
+      "epoch": 0.03680981595092025,
+      "grad_norm": 45.21313329781316,
+      "learning_rate": 7.983471074380164e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -334.0,
+      "logps/rejected": -266.0,
+      "loss": 0.4322,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 2.03125,
+      "rewards/rejected": -0.93359375,
+      "step": 30
+    },
+    {
+      "epoch": 0.03803680981595092,
+      "grad_norm": 41.258011933772664,
+      "learning_rate": 7.980165289256198e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -276.0,
+      "loss": 0.4115,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 2.765625,
+      "rewards/rejected": -1.1171875,
+      "step": 31
+    },
+    {
+      "epoch": 0.0392638036809816,
+      "grad_norm": 39.16160748489701,
+      "learning_rate": 7.976859504132231e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3836,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5859375,
+      "rewards/margins": 2.71875,
+      "rewards/rejected": -1.1328125,
+      "step": 32
+    },
+    {
+      "epoch": 0.04049079754601227,
+      "grad_norm": 38.84939653302124,
+      "learning_rate": 7.973553719008264e-07,
+      "logits/chosen": -0.0869140625,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -328.0,
+      "logps/rejected": -234.0,
+      "loss": 0.3966,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 2.234375,
+      "rewards/rejected": -1.0859375,
+      "step": 33
+    },
+    {
+      "epoch": 0.04171779141104295,
+      "grad_norm": 44.221302785049176,
+      "learning_rate": 7.970247933884297e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -302.0,
+      "loss": 0.4038,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 2.859375,
+      "rewards/rejected": -1.46875,
+      "step": 34
+    },
+    {
+      "epoch": 0.04294478527607362,
+      "grad_norm": 46.59951208776938,
+      "learning_rate": 7.96694214876033e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -264.0,
+      "loss": 0.3758,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 2.71875,
+      "rewards/rejected": -1.6484375,
+      "step": 35
+    },
+    {
+      "epoch": 0.044171779141104296,
+      "grad_norm": 41.317139252172254,
+      "learning_rate": 7.963636363636364e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3557,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 3.234375,
+      "rewards/rejected": -1.7421875,
+      "step": 36
+    },
+    {
+      "epoch": 0.04539877300613497,
+      "grad_norm": 39.795000825479654,
+      "learning_rate": 7.960330578512397e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3698,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 3.328125,
+      "rewards/rejected": -1.53125,
+      "step": 37
+    },
+    {
+      "epoch": 0.046625766871165646,
+      "grad_norm": 35.75368356475873,
+      "learning_rate": 7.95702479338843e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3582,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 3.6875,
+      "rewards/rejected": -1.671875,
+      "step": 38
+    },
+    {
+      "epoch": 0.04785276073619632,
+      "grad_norm": 46.53573926538421,
+      "learning_rate": 7.953719008264462e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3835,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 3.671875,
+      "rewards/rejected": -1.984375,
+      "step": 39
+    },
+    {
+      "epoch": 0.049079754601226995,
+      "grad_norm": 33.56284839867345,
+      "learning_rate": 7.950413223140495e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3042,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 3.765625,
+      "rewards/rejected": -2.109375,
+      "step": 40
+    },
+    {
+      "epoch": 0.05030674846625767,
+      "grad_norm": 39.46235435607716,
+      "learning_rate": 7.947107438016528e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -278.0,
+      "loss": 0.3708,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 3.546875,
+      "rewards/rejected": -2.0,
+      "step": 41
+    },
+    {
+      "epoch": 0.051533742331288344,
+      "grad_norm": 37.85600478708204,
+      "learning_rate": 7.943801652892562e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3236,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 3.59375,
+      "rewards/rejected": -2.15625,
+      "step": 42
+    },
+    {
+      "epoch": 0.05276073619631902,
+      "grad_norm": 43.531972474165585,
+      "learning_rate": 7.940495867768595e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3117,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6484375,
+      "rewards/margins": 4.0625,
+      "rewards/rejected": -2.421875,
+      "step": 43
+    },
+    {
+      "epoch": 0.053987730061349694,
+      "grad_norm": 63.54372274745974,
+      "learning_rate": 7.937190082644628e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -300.0,
+      "loss": 0.4265,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 3.625,
+      "rewards/rejected": -2.53125,
+      "step": 44
+    },
+    {
+      "epoch": 0.05521472392638037,
+      "grad_norm": 42.36111410389095,
+      "learning_rate": 7.933884297520661e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3513,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 3.875,
+      "rewards/rejected": -2.640625,
+      "step": 45
+    },
+    {
+      "epoch": 0.05644171779141104,
+      "grad_norm": 44.51420354088873,
+      "learning_rate": 7.930578512396694e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3109,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -2.609375,
+      "step": 46
+    },
+    {
+      "epoch": 0.05766871165644172,
+      "grad_norm": 61.22394416060924,
+      "learning_rate": 7.927272727272727e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -368.0,
+      "loss": 0.4432,
+      "rewards/accuracies": 0.7265625,
+      "rewards/chosen": 1.2109375,
+      "rewards/margins": 3.5,
+      "rewards/rejected": -2.28125,
+      "step": 47
+    },
+    {
+      "epoch": 0.05889570552147239,
+      "grad_norm": 35.37901763029379,
+      "learning_rate": 7.92396694214876e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -352.0,
+      "logps/rejected": -282.0,
+      "loss": 0.3719,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 3.203125,
+      "rewards/rejected": -2.09375,
+      "step": 48
+    },
+    {
+      "epoch": 0.06012269938650307,
+      "grad_norm": 31.606470718574474,
+      "learning_rate": 7.920661157024793e-07,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2642,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.90625,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.75,
+      "step": 49
+    },
+    {
+      "epoch": 0.06134969325153374,
+      "grad_norm": 48.38174059914219,
+      "learning_rate": 7.917355371900826e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3848,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 3.8125,
+      "rewards/rejected": -2.359375,
+      "step": 50
+    },
+    {
+      "epoch": 0.06257668711656442,
+      "grad_norm": 41.76003291409928,
+      "learning_rate": 7.914049586776859e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -442.0,
+      "logps/rejected": -358.0,
+      "loss": 0.2936,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.890625,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.640625,
+      "step": 51
+    },
+    {
+      "epoch": 0.0638036809815951,
+      "grad_norm": 36.258140724486836,
+      "learning_rate": 7.910743801652892e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3238,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 4.21875,
+      "rewards/rejected": -2.75,
+      "step": 52
+    },
+    {
+      "epoch": 0.06503067484662577,
+      "grad_norm": 38.82725433032811,
+      "learning_rate": 7.907438016528925e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3104,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 4.03125,
+      "rewards/rejected": -2.28125,
+      "step": 53
+    },
+    {
+      "epoch": 0.06625766871165645,
+      "grad_norm": 34.79397739539489,
+      "learning_rate": 7.904132231404959e-07,
+      "logits/chosen": -0.0284423828125,
+      "logits/rejected": -0.208984375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -258.0,
+      "loss": 0.2634,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 3.78125,
+      "rewards/rejected": -2.546875,
+      "step": 54
+    },
+    {
+      "epoch": 0.06748466257668712,
+      "grad_norm": 41.093192913712485,
+      "learning_rate": 7.900826446280992e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3081,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.515625,
+      "step": 55
+    },
+    {
+      "epoch": 0.0687116564417178,
+      "grad_norm": 45.49364963721349,
+      "learning_rate": 7.897520661157025e-07,
+      "logits/chosen": -0.119140625,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3753,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 4.0625,
+      "rewards/rejected": -2.265625,
+      "step": 56
+    },
+    {
+      "epoch": 0.06993865030674846,
+      "grad_norm": 50.666379541833365,
+      "learning_rate": 7.894214876033057e-07,
+      "logits/chosen": -0.1181640625,
+      "logits/rejected": -0.203125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3742,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 3.953125,
+      "rewards/rejected": -2.421875,
+      "step": 57
+    },
+    {
+      "epoch": 0.07116564417177915,
+      "grad_norm": 47.64145593969582,
+      "learning_rate": 7.89090909090909e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.2177734375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3598,
+      "rewards/accuracies": 0.7421875,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 4.125,
+      "rewards/rejected": -2.421875,
+      "step": 58
+    },
+    {
+      "epoch": 0.07239263803680981,
+      "grad_norm": 51.11345473280857,
+      "learning_rate": 7.887603305785123e-07,
+      "logits/chosen": -0.049560546875,
+      "logits/rejected": -0.140625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -292.0,
+      "loss": 0.4133,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 3.28125,
+      "rewards/rejected": -1.9921875,
+      "step": 59
+    },
+    {
+      "epoch": 0.0736196319018405,
+      "grad_norm": 39.0647396184868,
+      "learning_rate": 7.884297520661157e-07,
+      "logits/chosen": -0.12451171875,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -284.0,
+      "loss": 0.3578,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 3.46875,
+      "rewards/rejected": -2.28125,
+      "step": 60
+    },
+    {
+      "epoch": 0.07484662576687116,
+      "grad_norm": 39.12713457741509,
+      "learning_rate": 7.88099173553719e-07,
+      "logits/chosen": -0.125,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3224,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.953125,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.578125,
+      "step": 61
+    },
+    {
+      "epoch": 0.07607361963190185,
+      "grad_norm": 41.544347269900555,
+      "learning_rate": 7.877685950413223e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -428.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3433,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 2.328125,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -2.703125,
+      "step": 62
+    },
+    {
+      "epoch": 0.07730061349693251,
+      "grad_norm": 52.28530938487766,
+      "learning_rate": 7.874380165289256e-07,
+      "logits/chosen": -0.0257568359375,
+      "logits/rejected": -0.197265625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -304.0,
+      "loss": 0.331,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 3.9375,
+      "rewards/rejected": -2.40625,
+      "step": 63
+    },
+    {
+      "epoch": 0.0785276073619632,
+      "grad_norm": 44.343811483698694,
+      "learning_rate": 7.871074380165289e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3628,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -2.5,
+      "step": 64
+    },
+    {
+      "epoch": 0.07975460122699386,
+      "grad_norm": 40.38168553746162,
+      "learning_rate": 7.867768595041322e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3193,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.828125,
+      "rewards/margins": 4.28125,
+      "rewards/rejected": -2.453125,
+      "step": 65
+    },
+    {
+      "epoch": 0.08098159509202454,
+      "grad_norm": 35.90018196646264,
+      "learning_rate": 7.864462809917355e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -318.0,
+      "logps/rejected": -260.0,
+      "loss": 0.2892,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.0,
+      "rewards/margins": 4.8125,
+      "rewards/rejected": -2.8125,
+      "step": 66
+    },
+    {
+      "epoch": 0.08220858895705521,
+      "grad_norm": 39.83708858843296,
+      "learning_rate": 7.861157024793388e-07,
+      "logits/chosen": -0.1005859375,
+      "logits/rejected": -0.224609375,
+      "logps/chosen": -332.0,
+      "logps/rejected": -294.0,
+      "loss": 0.344,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 3.90625,
+      "rewards/rejected": -2.671875,
+      "step": 67
+    },
+    {
+      "epoch": 0.0834355828220859,
+      "grad_norm": 41.25508984701749,
+      "learning_rate": 7.857851239669421e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3597,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 3.953125,
+      "rewards/rejected": -2.46875,
+      "step": 68
+    },
+    {
+      "epoch": 0.08466257668711656,
+      "grad_norm": 35.61462182457114,
+      "learning_rate": 7.854545454545454e-07,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -282.0,
+      "loss": 0.2797,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.3125,
+      "rewards/margins": 4.96875,
+      "rewards/rejected": -2.65625,
+      "step": 69
+    },
+    {
+      "epoch": 0.08588957055214724,
+      "grad_norm": 43.949799113654095,
+      "learning_rate": 7.851239669421487e-07,
+      "logits/chosen": -0.12109375,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3016,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.328125,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.34375,
+      "step": 70
+    },
+    {
+      "epoch": 0.08711656441717791,
+      "grad_norm": 40.385403541291616,
+      "learning_rate": 7.84793388429752e-07,
+      "logits/chosen": -0.068359375,
+      "logits/rejected": -0.1494140625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -278.0,
+      "loss": 0.3925,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 3.625,
+      "rewards/rejected": -2.265625,
+      "step": 71
+    },
+    {
+      "epoch": 0.08834355828220859,
+      "grad_norm": 42.66471807743761,
+      "learning_rate": 7.844628099173554e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -430.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2955,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9609375,
+      "rewards/margins": 4.25,
+      "rewards/rejected": -2.28125,
+      "step": 72
+    },
+    {
+      "epoch": 0.08957055214723926,
+      "grad_norm": 45.57465316917817,
+      "learning_rate": 7.841322314049587e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -348.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3924,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.5625,
+      "step": 73
+    },
+    {
+      "epoch": 0.09079754601226994,
+      "grad_norm": 41.41678309244041,
+      "learning_rate": 7.838016528925619e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2911,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 3.9375,
+      "rewards/rejected": -2.65625,
+      "step": 74
+    },
+    {
+      "epoch": 0.09202453987730061,
+      "grad_norm": 36.36740264063768,
+      "learning_rate": 7.834710743801652e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -318.0,
+      "loss": 0.271,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.703125,
+      "step": 75
+    },
+    {
+      "epoch": 0.09325153374233129,
+      "grad_norm": 51.152036733559505,
+      "learning_rate": 7.831404958677685e-07,
+      "logits/chosen": -0.1513671875,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -288.0,
+      "loss": 0.3262,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -2.53125,
+      "step": 76
+    },
+    {
+      "epoch": 0.09447852760736196,
+      "grad_norm": 34.55182641721203,
+      "learning_rate": 7.828099173553718e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2774,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.4453125,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -2.875,
+      "step": 77
+    },
+    {
+      "epoch": 0.09570552147239264,
+      "grad_norm": 40.758007342891354,
+      "learning_rate": 7.824793388429752e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2953,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -2.859375,
+      "step": 78
+    },
+    {
+      "epoch": 0.09693251533742331,
+      "grad_norm": 40.243771187807226,
+      "learning_rate": 7.821487603305785e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -286.0,
+      "loss": 0.324,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.796875,
+      "step": 79
+    },
+    {
+      "epoch": 0.09815950920245399,
+      "grad_norm": 37.86169713828253,
+      "learning_rate": 7.818181818181818e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2795,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.0,
+      "step": 80
+    },
+    {
+      "epoch": 0.09938650306748466,
+      "grad_norm": 36.69542578194377,
+      "learning_rate": 7.814876033057851e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -436.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.9375,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.09375,
+      "step": 81
+    },
+    {
+      "epoch": 0.10061349693251534,
+      "grad_norm": 52.318975452049024,
+      "learning_rate": 7.811570247933884e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3562,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -3.1875,
+      "step": 82
+    },
+    {
+      "epoch": 0.10184049079754601,
+      "grad_norm": 48.36250064527517,
+      "learning_rate": 7.808264462809916e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3065,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 4.5625,
+      "rewards/rejected": -3.234375,
+      "step": 83
+    },
+    {
+      "epoch": 0.10306748466257669,
+      "grad_norm": 56.4532578365936,
+      "learning_rate": 7.80495867768595e-07,
+      "logits/chosen": -0.11083984375,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -448.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2942,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -3.53125,
+      "step": 84
+    },
+    {
+      "epoch": 0.10429447852760736,
+      "grad_norm": 40.84421229525563,
+      "learning_rate": 7.801652892561983e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3021,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.484375,
+      "step": 85
+    },
+    {
+      "epoch": 0.10552147239263804,
+      "grad_norm": 42.8824866796569,
+      "learning_rate": 7.798347107438016e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -302.0,
+      "loss": 0.3346,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.03125,
+      "step": 86
+    },
+    {
+      "epoch": 0.1067484662576687,
+      "grad_norm": 32.177789360920215,
+      "learning_rate": 7.795041322314049e-07,
+      "logits/chosen": -0.17578125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -260.0,
+      "loss": 0.2835,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 4.25,
+      "rewards/rejected": -3.46875,
+      "step": 87
+    },
+    {
+      "epoch": 0.10797546012269939,
+      "grad_norm": 34.28909140331474,
+      "learning_rate": 7.791735537190082e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -276.0,
+      "loss": 0.265,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -3.3125,
+      "step": 88
+    },
+    {
+      "epoch": 0.10920245398773006,
+      "grad_norm": 42.40518360379319,
+      "learning_rate": 7.788429752066116e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3464,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.4140625,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -3.09375,
+      "step": 89
+    },
+    {
+      "epoch": 0.11042944785276074,
+      "grad_norm": 49.53932371851528,
+      "learning_rate": 7.785123966942149e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3253,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.98828125,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -3.3125,
+      "step": 90
+    },
+    {
+      "epoch": 0.1116564417177914,
+      "grad_norm": 40.48862678893552,
+      "learning_rate": 7.781818181818182e-07,
+      "logits/chosen": -0.039794921875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2845,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -3.203125,
+      "step": 91
+    },
+    {
+      "epoch": 0.11288343558282209,
+      "grad_norm": 44.87681376045074,
+      "learning_rate": 7.778512396694214e-07,
+      "logits/chosen": -0.126953125,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2738,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.5,
+      "step": 92
+    },
+    {
+      "epoch": 0.11411042944785275,
+      "grad_norm": 36.61224632474363,
+      "learning_rate": 7.775206611570247e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -300.0,
+      "loss": 0.282,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.3125,
+      "step": 93
+    },
+    {
+      "epoch": 0.11533742331288344,
+      "grad_norm": 42.24461574764279,
+      "learning_rate": 7.77190082644628e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2813,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.234375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.140625,
+      "step": 94
+    },
+    {
+      "epoch": 0.1165644171779141,
+      "grad_norm": 42.13498046583409,
+      "learning_rate": 7.768595041322314e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.7734375,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -3.390625,
+      "step": 95
+    },
+    {
+      "epoch": 0.11779141104294479,
+      "grad_norm": 39.13305361039425,
+      "learning_rate": 7.765289256198347e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -310.0,
+      "loss": 0.253,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.203125,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -3.40625,
+      "step": 96
+    },
+    {
+      "epoch": 0.11901840490797547,
+      "grad_norm": 46.1275059887527,
+      "learning_rate": 7.76198347107438e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -352.0,
+      "loss": 0.357,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -2.890625,
+      "step": 97
+    },
+    {
+      "epoch": 0.12024539877300613,
+      "grad_norm": 38.580559184520915,
+      "learning_rate": 7.758677685950413e-07,
+      "logits/chosen": -0.041015625,
+      "logits/rejected": -0.248046875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -282.0,
+      "loss": 0.3015,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.828125,
+      "step": 98
+    },
+    {
+      "epoch": 0.12147239263803682,
+      "grad_norm": 38.975117366609354,
+      "learning_rate": 7.755371900826446e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -340.0,
+      "logps/rejected": -244.0,
+      "loss": 0.2586,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.84375,
+      "step": 99
+    },
+    {
+      "epoch": 0.12269938650306748,
+      "grad_norm": 38.2314125667712,
+      "learning_rate": 7.752066115702479e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3048,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 4.875,
+      "rewards/rejected": -3.09375,
+      "step": 100
+    },
+    {
+      "epoch": 0.12392638036809817,
+      "grad_norm": 47.72088602593496,
+      "learning_rate": 7.748760330578512e-07,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3549,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.921875,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -2.984375,
+      "step": 101
+    },
+    {
+      "epoch": 0.12515337423312883,
+      "grad_norm": 39.20874470259191,
+      "learning_rate": 7.745454545454545e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3073,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.9375,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.71875,
+      "step": 102
+    },
+    {
+      "epoch": 0.1263803680981595,
+      "grad_norm": 36.963577318972284,
+      "learning_rate": 7.742148760330578e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2878,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7265625,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -2.75,
+      "step": 103
+    },
+    {
+      "epoch": 0.1276073619631902,
+      "grad_norm": 50.89003123288155,
+      "learning_rate": 7.738842975206611e-07,
+      "logits/chosen": -0.107421875,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3679,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 4.40625,
+      "rewards/rejected": -2.71875,
+      "step": 104
+    },
+    {
+      "epoch": 0.12883435582822086,
+      "grad_norm": 34.19259318441508,
+      "learning_rate": 7.735537190082644e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2586,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.984375,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.140625,
+      "step": 105
+    },
+    {
+      "epoch": 0.13006134969325153,
+      "grad_norm": 38.48374397973381,
+      "learning_rate": 7.732231404958677e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3111,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.9140625,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -2.828125,
+      "step": 106
+    },
+    {
+      "epoch": 0.1312883435582822,
+      "grad_norm": 41.75637341421824,
+      "learning_rate": 7.728925619834711e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2869,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.84375,
+      "step": 107
+    },
+    {
+      "epoch": 0.1325153374233129,
+      "grad_norm": 33.954750583294235,
+      "learning_rate": 7.725619834710744e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2625,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.328125,
+      "step": 108
+    },
+    {
+      "epoch": 0.13374233128834356,
+      "grad_norm": 36.44619320062165,
+      "learning_rate": 7.722314049586777e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2958,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.6875,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -2.96875,
+      "step": 109
+    },
+    {
+      "epoch": 0.13496932515337423,
+      "grad_norm": 41.465536696868014,
+      "learning_rate": 7.719008264462809e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3213,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 2.234375,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.015625,
+      "step": 110
+    },
+    {
+      "epoch": 0.1361963190184049,
+      "grad_norm": 34.97820982991418,
+      "learning_rate": 7.715702479338842e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -330.0,
+      "logps/rejected": -268.0,
+      "loss": 0.2777,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 4.40625,
+      "rewards/rejected": -2.75,
+      "step": 111
+    },
+    {
+      "epoch": 0.1374233128834356,
+      "grad_norm": 37.33233370442031,
+      "learning_rate": 7.712396694214875e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -340.0,
+      "logps/rejected": -278.0,
+      "loss": 0.2729,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 4.6875,
+      "rewards/rejected": -3.125,
+      "step": 112
+    },
+    {
+      "epoch": 0.13865030674846626,
+      "grad_norm": 39.32831836749558,
+      "learning_rate": 7.709090909090909e-07,
+      "logits/chosen": -0.1044921875,
+      "logits/rejected": -0.2001953125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -302.0,
+      "loss": 0.3238,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.84375,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -2.890625,
+      "step": 113
+    },
+    {
+      "epoch": 0.13987730061349693,
+      "grad_norm": 43.77160490529073,
+      "learning_rate": 7.705785123966942e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3356,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -2.96875,
+      "step": 114
+    },
+    {
+      "epoch": 0.1411042944785276,
+      "grad_norm": 45.86173030028754,
+      "learning_rate": 7.702479338842975e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3666,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 4.0,
+      "rewards/rejected": -2.96875,
+      "step": 115
+    },
+    {
+      "epoch": 0.1423312883435583,
+      "grad_norm": 31.833834890400677,
+      "learning_rate": 7.699173553719008e-07,
+      "logits/chosen": -0.2001953125,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2561,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.078125,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.140625,
+      "step": 116
+    },
+    {
+      "epoch": 0.14355828220858896,
+      "grad_norm": 34.65342816309822,
+      "learning_rate": 7.695867768595041e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -324.0,
+      "logps/rejected": -262.0,
+      "loss": 0.2802,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -3.453125,
+      "step": 117
+    },
+    {
+      "epoch": 0.14478527607361963,
+      "grad_norm": 40.64680750199124,
+      "learning_rate": 7.692561983471075e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3232,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.09375,
+      "step": 118
+    },
+    {
+      "epoch": 0.1460122699386503,
+      "grad_norm": 39.06542225750342,
+      "learning_rate": 7.689256198347107e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3009,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -3.34375,
+      "step": 119
+    },
+    {
+      "epoch": 0.147239263803681,
+      "grad_norm": 45.793161618342445,
+      "learning_rate": 7.68595041322314e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2967,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 2.421875,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -3.234375,
+      "step": 120
+    },
+    {
+      "epoch": 0.14846625766871166,
+      "grad_norm": 36.35506762175528,
+      "learning_rate": 7.682644628099173e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2827,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 2.265625,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -3.09375,
+      "step": 121
+    },
+    {
+      "epoch": 0.14969325153374233,
+      "grad_norm": 43.61172830028863,
+      "learning_rate": 7.679338842975206e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2954,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -3.015625,
+      "step": 122
+    },
+    {
+      "epoch": 0.150920245398773,
+      "grad_norm": 44.76510892093205,
+      "learning_rate": 7.676033057851239e-07,
+      "logits/chosen": -0.054443359375,
+      "logits/rejected": -0.2294921875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -278.0,
+      "loss": 0.3266,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 4.40625,
+      "rewards/rejected": -3.171875,
+      "step": 123
+    },
+    {
+      "epoch": 0.1521472392638037,
+      "grad_norm": 34.74558330134135,
+      "learning_rate": 7.672727272727273e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -350.0,
+      "logps/rejected": -280.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -3.46875,
+      "step": 124
+    },
+    {
+      "epoch": 0.15337423312883436,
+      "grad_norm": 42.40382525839625,
+      "learning_rate": 7.669421487603306e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2816,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.765625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.5625,
+      "step": 125
+    },
+    {
+      "epoch": 0.15460122699386503,
+      "grad_norm": 54.09847924655591,
+      "learning_rate": 7.666115702479339e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -302.0,
+      "loss": 0.3503,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 4.34375,
+      "rewards/rejected": -3.34375,
+      "step": 126
+    },
+    {
+      "epoch": 0.1558282208588957,
+      "grad_norm": 42.126209906017394,
+      "learning_rate": 7.662809917355372e-07,
+      "logits/chosen": -0.09375,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3524,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.96875,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.15625,
+      "step": 127
+    },
+    {
+      "epoch": 0.1570552147239264,
+      "grad_norm": 41.454801423663085,
+      "learning_rate": 7.659504132231404e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2841,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8671875,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.390625,
+      "step": 128
+    },
+    {
+      "epoch": 0.15828220858895706,
+      "grad_norm": 37.48031578456685,
+      "learning_rate": 7.656198347107437e-07,
+      "logits/chosen": -0.0478515625,
+      "logits/rejected": -0.1796875,
+      "logps/chosen": -336.0,
+      "logps/rejected": -284.0,
+      "loss": 0.3245,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -3.0,
+      "step": 129
+    },
+    {
+      "epoch": 0.15950920245398773,
+      "grad_norm": 38.14450880210714,
+      "learning_rate": 7.65289256198347e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3273,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -3.421875,
+      "step": 130
+    },
+    {
+      "epoch": 0.1607361963190184,
+      "grad_norm": 35.19502262660919,
+      "learning_rate": 7.649586776859504e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2106,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.953125,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.59375,
+      "step": 131
+    },
+    {
+      "epoch": 0.1619631901840491,
+      "grad_norm": 42.07184086801236,
+      "learning_rate": 7.646280991735537e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3252,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.765625,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -3.40625,
+      "step": 132
+    },
+    {
+      "epoch": 0.16319018404907976,
+      "grad_norm": 43.059635351413924,
+      "learning_rate": 7.64297520661157e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2872,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.703125,
+      "step": 133
+    },
+    {
+      "epoch": 0.16441717791411042,
+      "grad_norm": 48.09563853008669,
+      "learning_rate": 7.639669421487603e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -450.0,
+      "logps/rejected": -358.0,
+      "loss": 0.2946,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 4.625,
+      "rewards/rejected": -3.109375,
+      "step": 134
+    },
+    {
+      "epoch": 0.1656441717791411,
+      "grad_norm": 46.36508764131004,
+      "learning_rate": 7.636363636363636e-07,
+      "logits/chosen": -0.05712890625,
+      "logits/rejected": -0.171875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3654,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.9765625,
+      "rewards/margins": 4.375,
+      "rewards/rejected": -3.390625,
+      "step": 135
+    },
+    {
+      "epoch": 0.1668711656441718,
+      "grad_norm": 39.349987951574626,
+      "learning_rate": 7.63305785123967e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3195,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -3.78125,
+      "step": 136
+    },
+    {
+      "epoch": 0.16809815950920245,
+      "grad_norm": 46.78307320347251,
+      "learning_rate": 7.629752066115702e-07,
+      "logits/chosen": -0.061279296875,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3621,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 4.25,
+      "rewards/rejected": -3.078125,
+      "step": 137
+    },
+    {
+      "epoch": 0.16932515337423312,
+      "grad_norm": 37.18351460769713,
+      "learning_rate": 7.626446280991735e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3022,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.46875,
+      "step": 138
+    },
+    {
+      "epoch": 0.1705521472392638,
+      "grad_norm": 42.545002180919454,
+      "learning_rate": 7.623140495867768e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -362.0,
+      "loss": 0.352,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -3.828125,
+      "step": 139
+    },
+    {
+      "epoch": 0.17177914110429449,
+      "grad_norm": 47.45496980791103,
+      "learning_rate": 7.619834710743801e-07,
+      "logits/chosen": -0.10400390625,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3166,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.5625,
+      "step": 140
+    },
+    {
+      "epoch": 0.17300613496932515,
+      "grad_norm": 34.622965144928166,
+      "learning_rate": 7.616528925619834e-07,
+      "logits/chosen": -0.0849609375,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.251,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.515625,
+      "step": 141
+    },
+    {
+      "epoch": 0.17423312883435582,
+      "grad_norm": 40.37071693879945,
+      "learning_rate": 7.613223140495868e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3233,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.640625,
+      "step": 142
+    },
+    {
+      "epoch": 0.1754601226993865,
+      "grad_norm": 41.48963188375289,
+      "learning_rate": 7.609917355371901e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2941,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.828125,
+      "step": 143
+    },
+    {
+      "epoch": 0.17668711656441718,
+      "grad_norm": 40.697834722668915,
+      "learning_rate": 7.606611570247934e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2905,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.96875,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.59375,
+      "step": 144
+    },
+    {
+      "epoch": 0.17791411042944785,
+      "grad_norm": 39.56968971060715,
+      "learning_rate": 7.603305785123967e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2715,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.65625,
+      "step": 145
+    },
+    {
+      "epoch": 0.17914110429447852,
+      "grad_norm": 34.52865223755899,
+      "learning_rate": 7.599999999999999e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2436,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.78125,
+      "step": 146
+    },
+    {
+      "epoch": 0.18036809815950922,
+      "grad_norm": 35.86987627703754,
+      "learning_rate": 7.596694214876032e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2741,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.84375,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.71875,
+      "step": 147
+    },
+    {
+      "epoch": 0.18159509202453988,
+      "grad_norm": 49.681572441729514,
+      "learning_rate": 7.593388429752066e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3761,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.625,
+      "step": 148
+    },
+    {
+      "epoch": 0.18282208588957055,
+      "grad_norm": 43.932202256712266,
+      "learning_rate": 7.590082644628099e-07,
+      "logits/chosen": -0.11474609375,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2534,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.984375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.390625,
+      "step": 149
+    },
+    {
+      "epoch": 0.18404907975460122,
+      "grad_norm": 39.402366627439726,
+      "learning_rate": 7.586776859504132e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3158,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -3.09375,
+      "step": 150
+    },
+    {
+      "epoch": 0.18527607361963191,
+      "grad_norm": 41.79252221918161,
+      "learning_rate": 7.583471074380165e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3211,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -3.03125,
+      "step": 151
+    },
+    {
+      "epoch": 0.18650306748466258,
+      "grad_norm": 35.715508495832665,
+      "learning_rate": 7.580165289256198e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -298.0,
+      "loss": 0.2622,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.921875,
+      "step": 152
+    },
+    {
+      "epoch": 0.18773006134969325,
+      "grad_norm": 36.417969122212845,
+      "learning_rate": 7.576859504132231e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2871,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.625,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.390625,
+      "step": 153
+    },
+    {
+      "epoch": 0.18895705521472392,
+      "grad_norm": 45.428337431051204,
+      "learning_rate": 7.573553719008265e-07,
+      "logits/chosen": -0.09716796875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3165,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 2.046875,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -2.984375,
+      "step": 154
+    },
+    {
+      "epoch": 0.1901840490797546,
+      "grad_norm": 28.34310543484037,
+      "learning_rate": 7.570247933884297e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1864,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 2.5625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -3.65625,
+      "step": 155
+    },
+    {
+      "epoch": 0.19141104294478528,
+      "grad_norm": 43.05423041815277,
+      "learning_rate": 7.56694214876033e-07,
+      "logits/chosen": -0.11865234375,
+      "logits/rejected": -0.2314453125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2995,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 4.5625,
+      "rewards/rejected": -3.09375,
+      "step": 156
+    },
+    {
+      "epoch": 0.19263803680981595,
+      "grad_norm": 35.31331641855709,
+      "learning_rate": 7.563636363636363e-07,
+      "logits/chosen": -0.11474609375,
+      "logits/rejected": -0.244140625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2791,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 4.875,
+      "rewards/rejected": -3.328125,
+      "step": 157
+    },
+    {
+      "epoch": 0.19386503067484662,
+      "grad_norm": 32.52885612180771,
+      "learning_rate": 7.560330578512396e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2615,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.6875,
+      "step": 158
+    },
+    {
+      "epoch": 0.1950920245398773,
+      "grad_norm": 40.74250145862389,
+      "learning_rate": 7.55702479338843e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -318.0,
+      "loss": 0.312,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.25,
+      "step": 159
+    },
+    {
+      "epoch": 0.19631901840490798,
+      "grad_norm": 32.03909927740134,
+      "learning_rate": 7.553719008264463e-07,
+      "logits/chosen": -0.10009765625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2149,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.765625,
+      "step": 160
+    },
+    {
+      "epoch": 0.19754601226993865,
+      "grad_norm": 33.48229976640365,
+      "learning_rate": 7.550413223140496e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2313,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.8984375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -3.703125,
+      "step": 161
+    },
+    {
+      "epoch": 0.19877300613496932,
+      "grad_norm": 44.9989690142436,
+      "learning_rate": 7.547107438016529e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3547,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.390625,
+      "step": 162
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 45.392211600179095,
+      "learning_rate": 7.543801652892561e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3203,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.4375,
+      "step": 163
+    },
+    {
+      "epoch": 0.20122699386503068,
+      "grad_norm": 30.10144812491185,
+      "learning_rate": 7.540495867768594e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -298.0,
+      "loss": 0.2205,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -3.953125,
+      "step": 164
+    },
+    {
+      "epoch": 0.20245398773006135,
+      "grad_norm": 41.70055038557836,
+      "learning_rate": 7.537190082644627e-07,
+      "logits/chosen": -0.078125,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -440.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2684,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.7578125,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.484375,
+      "step": 165
+    },
+    {
+      "epoch": 0.20368098159509201,
+      "grad_norm": 40.19134249676154,
+      "learning_rate": 7.533884297520661e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3382,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -3.296875,
+      "step": 166
+    },
+    {
+      "epoch": 0.2049079754601227,
+      "grad_norm": 40.73221481198693,
+      "learning_rate": 7.530578512396694e-07,
+      "logits/chosen": -0.1240234375,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3475,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 2.21875,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -3.40625,
+      "step": 167
+    },
+    {
+      "epoch": 0.20613496932515338,
+      "grad_norm": 36.56704570459498,
+      "learning_rate": 7.527272727272727e-07,
+      "logits/chosen": -0.10888671875,
+      "logits/rejected": -0.2236328125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -276.0,
+      "loss": 0.2594,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.0,
+      "step": 168
+    },
+    {
+      "epoch": 0.20736196319018405,
+      "grad_norm": 36.86789794879476,
+      "learning_rate": 7.52396694214876e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2793,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.75,
+      "step": 169
+    },
+    {
+      "epoch": 0.2085889570552147,
+      "grad_norm": 43.700977925857806,
+      "learning_rate": 7.520661157024793e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -448.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2994,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.046875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -3.765625,
+      "step": 170
+    },
+    {
+      "epoch": 0.2098159509202454,
+      "grad_norm": 33.49786387415663,
+      "learning_rate": 7.517355371900827e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2804,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.7734375,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.0625,
+      "step": 171
+    },
+    {
+      "epoch": 0.21104294478527608,
+      "grad_norm": 36.775839118353844,
+      "learning_rate": 7.514049586776859e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -436.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2514,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.03125,
+      "step": 172
+    },
+    {
+      "epoch": 0.21226993865030674,
+      "grad_norm": 29.010747243935523,
+      "learning_rate": 7.510743801652892e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2075,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.296875,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.125,
+      "step": 173
+    },
+    {
+      "epoch": 0.2134969325153374,
+      "grad_norm": 42.10282379400345,
+      "learning_rate": 7.507438016528925e-07,
+      "logits/chosen": -0.035400390625,
+      "logits/rejected": -0.248046875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3371,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.890625,
+      "step": 174
+    },
+    {
+      "epoch": 0.2147239263803681,
+      "grad_norm": 44.4053310320228,
+      "learning_rate": 7.504132231404958e-07,
+      "logits/chosen": -0.17578125,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3132,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.9609375,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -3.921875,
+      "step": 175
+    },
+    {
+      "epoch": 0.21595092024539878,
+      "grad_norm": 38.79125305740806,
+      "learning_rate": 7.500826446280991e-07,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -354.0,
+      "loss": 0.315,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -3.875,
+      "step": 176
+    },
+    {
+      "epoch": 0.21717791411042944,
+      "grad_norm": 29.08349894872895,
+      "learning_rate": 7.497520661157025e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2249,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.453125,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -4.375,
+      "step": 177
+    },
+    {
+      "epoch": 0.2184049079754601,
+      "grad_norm": 39.84454997153339,
+      "learning_rate": 7.494214876033058e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3147,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.375,
+      "step": 178
+    },
+    {
+      "epoch": 0.2196319018404908,
+      "grad_norm": 32.49552053062871,
+      "learning_rate": 7.490909090909091e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2585,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.34375,
+      "step": 179
+    },
+    {
+      "epoch": 0.22085889570552147,
+      "grad_norm": 41.53796098392134,
+      "learning_rate": 7.487603305785124e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3355,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.75,
+      "step": 180
+    },
+    {
+      "epoch": 0.22208588957055214,
+      "grad_norm": 43.27113784190138,
+      "learning_rate": 7.484297520661156e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -372.0,
+      "loss": 0.3558,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.03125,
+      "step": 181
+    },
+    {
+      "epoch": 0.2233128834355828,
+      "grad_norm": 40.060092182372635,
+      "learning_rate": 7.480991735537189e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3329,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.92578125,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -4.0625,
+      "step": 182
+    },
+    {
+      "epoch": 0.2245398773006135,
+      "grad_norm": 42.893284410160845,
+      "learning_rate": 7.477685950413223e-07,
+      "logits/chosen": -0.0888671875,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -374.0,
+      "loss": 0.3303,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -3.703125,
+      "step": 183
+    },
+    {
+      "epoch": 0.22576687116564417,
+      "grad_norm": 44.81738203832481,
+      "learning_rate": 7.474380165289256e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.2431640625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3382,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.0625,
+      "step": 184
+    },
+    {
+      "epoch": 0.22699386503067484,
+      "grad_norm": 49.18179746490089,
+      "learning_rate": 7.471074380165289e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3425,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -3.734375,
+      "step": 185
+    },
+    {
+      "epoch": 0.2282208588957055,
+      "grad_norm": 38.62632364463619,
+      "learning_rate": 7.467768595041322e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3061,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.953125,
+      "step": 186
+    },
+    {
+      "epoch": 0.2294478527607362,
+      "grad_norm": 37.48238102625356,
+      "learning_rate": 7.464462809917355e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2719,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.375,
+      "step": 187
+    },
+    {
+      "epoch": 0.23067484662576687,
+      "grad_norm": 33.9965710800535,
+      "learning_rate": 7.461157024793388e-07,
+      "logits/chosen": -0.0712890625,
+      "logits/rejected": -0.1689453125,
+      "logps/chosen": -332.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2908,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.09375,
+      "step": 188
+    },
+    {
+      "epoch": 0.23190184049079754,
+      "grad_norm": 32.6251624460159,
+      "learning_rate": 7.457851239669422e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2196,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.3125,
+      "step": 189
+    },
+    {
+      "epoch": 0.2331288343558282,
+      "grad_norm": 33.86427029616883,
+      "learning_rate": 7.454545454545454e-07,
+      "logits/chosen": -0.05810546875,
+      "logits/rejected": -0.224609375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2859,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.875,
+      "step": 190
+    },
+    {
+      "epoch": 0.2343558282208589,
+      "grad_norm": 48.95754374693823,
+      "learning_rate": 7.451239669421487e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3495,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.1875,
+      "step": 191
+    },
+    {
+      "epoch": 0.23558282208588957,
+      "grad_norm": 44.694192367882906,
+      "learning_rate": 7.44793388429752e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -292.0,
+      "loss": 0.3014,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.578125,
+      "step": 192
+    },
+    {
+      "epoch": 0.23680981595092024,
+      "grad_norm": 34.25684662561766,
+      "learning_rate": 7.444628099173553e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2177,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.28125,
+      "step": 193
+    },
+    {
+      "epoch": 0.23803680981595093,
+      "grad_norm": 39.119509435238996,
+      "learning_rate": 7.441322314049586e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2953,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.4375,
+      "step": 194
+    },
+    {
+      "epoch": 0.2392638036809816,
+      "grad_norm": 34.817335787332766,
+      "learning_rate": 7.43801652892562e-07,
+      "logits/chosen": -0.1181640625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2301,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.4375,
+      "step": 195
+    },
+    {
+      "epoch": 0.24049079754601227,
+      "grad_norm": 40.102802913255154,
+      "learning_rate": 7.434710743801653e-07,
+      "logits/chosen": -0.09130859375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -290.0,
+      "loss": 0.297,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -3.828125,
+      "step": 196
+    },
+    {
+      "epoch": 0.24171779141104294,
+      "grad_norm": 34.798967486094014,
+      "learning_rate": 7.431404958677686e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2803,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.0625,
+      "step": 197
+    },
+    {
+      "epoch": 0.24294478527607363,
+      "grad_norm": 39.54376141056963,
+      "learning_rate": 7.428099173553719e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.2265625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -258.0,
+      "loss": 0.3347,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.671875,
+      "step": 198
+    },
+    {
+      "epoch": 0.2441717791411043,
+      "grad_norm": 29.961664099131877,
+      "learning_rate": 7.424793388429751e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -428.0,
+      "logps/rejected": -310.0,
+      "loss": 0.1903,
+      "rewards/accuracies": 0.9140625,
+      "rewards/chosen": 2.359375,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.09375,
+      "step": 199
+    },
+    {
+      "epoch": 0.24539877300613497,
+      "grad_norm": 41.86791512110987,
+      "learning_rate": 7.421487603305784e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -370.0,
+      "loss": 0.3056,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.203125,
+      "rewards/margins": 6.125,
+      "rewards/rejected": -3.921875,
+      "step": 200
+    },
+    {
+      "epoch": 0.24662576687116564,
+      "grad_norm": 32.0817049971399,
+      "learning_rate": 7.418181818181818e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2166,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 2.0625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -3.734375,
+      "step": 201
+    },
+    {
+      "epoch": 0.24785276073619633,
+      "grad_norm": 35.7245529395789,
+      "learning_rate": 7.414876033057851e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2641,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -3.640625,
+      "step": 202
+    },
+    {
+      "epoch": 0.249079754601227,
+      "grad_norm": 42.27175209208064,
+      "learning_rate": 7.411570247933884e-07,
+      "logits/chosen": -0.08935546875,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4055,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8828125,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.03125,
+      "step": 203
+    },
+    {
+      "epoch": 0.25030674846625767,
+      "grad_norm": 39.50994905620242,
+      "learning_rate": 7.408264462809917e-07,
+      "logits/chosen": -0.0771484375,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -268.0,
+      "loss": 0.3395,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -3.671875,
+      "step": 204
+    },
+    {
+      "epoch": 0.25153374233128833,
+      "grad_norm": 37.49332179023069,
+      "learning_rate": 7.40495867768595e-07,
+      "logits/chosen": -0.02880859375,
+      "logits/rejected": -0.173828125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3209,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.84375,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.265625,
+      "step": 205
+    },
+    {
+      "epoch": 0.252760736196319,
+      "grad_norm": 27.399492756017406,
+      "learning_rate": 7.401652892561984e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2051,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.28125,
+      "step": 206
+    },
+    {
+      "epoch": 0.25398773006134967,
+      "grad_norm": 43.41671912422527,
+      "learning_rate": 7.398347107438017e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3584,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 2.140625,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -3.71875,
+      "step": 207
+    },
+    {
+      "epoch": 0.2552147239263804,
+      "grad_norm": 38.06892744188352,
+      "learning_rate": 7.395041322314049e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3179,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.71875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.09375,
+      "step": 208
+    },
+    {
+      "epoch": 0.25644171779141106,
+      "grad_norm": 38.21598347000612,
+      "learning_rate": 7.391735537190082e-07,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -336.0,
+      "loss": 0.306,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -3.90625,
+      "step": 209
+    },
+    {
+      "epoch": 0.25766871165644173,
+      "grad_norm": 45.46016539899385,
+      "learning_rate": 7.388429752066115e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -364.0,
+      "loss": 0.3476,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.9375,
+      "step": 210
+    },
+    {
+      "epoch": 0.2588957055214724,
+      "grad_norm": 38.3264065355804,
+      "learning_rate": 7.385123966942148e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3035,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.53125,
+      "step": 211
+    },
+    {
+      "epoch": 0.26012269938650306,
+      "grad_norm": 37.36895578645757,
+      "learning_rate": 7.381818181818182e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2805,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.28125,
+      "step": 212
+    },
+    {
+      "epoch": 0.26134969325153373,
+      "grad_norm": 33.79330367194955,
+      "learning_rate": 7.378512396694215e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -284.0,
+      "loss": 0.2588,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.9375,
+      "step": 213
+    },
+    {
+      "epoch": 0.2625766871165644,
+      "grad_norm": 41.91194027673872,
+      "learning_rate": 7.375206611570248e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3085,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -4.125,
+      "step": 214
+    },
+    {
+      "epoch": 0.26380368098159507,
+      "grad_norm": 42.67216156371333,
+      "learning_rate": 7.371900826446281e-07,
+      "logits/chosen": -0.0927734375,
+      "logits/rejected": -0.1953125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3066,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 4.96875,
+      "rewards/rejected": -3.8125,
+      "step": 215
+    },
+    {
+      "epoch": 0.2650306748466258,
+      "grad_norm": 41.6917127385005,
+      "learning_rate": 7.368595041322314e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -330.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3077,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -3.734375,
+      "step": 216
+    },
+    {
+      "epoch": 0.26625766871165646,
+      "grad_norm": 41.999916191643734,
+      "learning_rate": 7.365289256198346e-07,
+      "logits/chosen": -0.05322265625,
+      "logits/rejected": -0.20703125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3621,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.421875,
+      "step": 217
+    },
+    {
+      "epoch": 0.2674846625766871,
+      "grad_norm": 48.35045291177634,
+      "learning_rate": 7.36198347107438e-07,
+      "logits/chosen": -0.10009765625,
+      "logits/rejected": -0.21875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3724,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 4.875,
+      "rewards/rejected": -3.515625,
+      "step": 218
+    },
+    {
+      "epoch": 0.2687116564417178,
+      "grad_norm": 36.99776737120032,
+      "learning_rate": 7.358677685950413e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3041,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.4140625,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -3.875,
+      "step": 219
+    },
+    {
+      "epoch": 0.26993865030674846,
+      "grad_norm": 49.4734461720144,
+      "learning_rate": 7.355371900826446e-07,
+      "logits/chosen": -0.05078125,
+      "logits/rejected": -0.1318359375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3781,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 4.375,
+      "rewards/rejected": -3.421875,
+      "step": 220
+    },
+    {
+      "epoch": 0.27116564417177913,
+      "grad_norm": 50.14968062203286,
+      "learning_rate": 7.352066115702479e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -292.0,
+      "loss": 0.3828,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.796875,
+      "step": 221
+    },
+    {
+      "epoch": 0.2723926380368098,
+      "grad_norm": 35.95574239692987,
+      "learning_rate": 7.348760330578512e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2649,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.09375,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.15625,
+      "step": 222
+    },
+    {
+      "epoch": 0.27361963190184047,
+      "grad_norm": 40.443295860960646,
+      "learning_rate": 7.345454545454545e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2888,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.75,
+      "step": 223
+    },
+    {
+      "epoch": 0.2748466257668712,
+      "grad_norm": 42.19474227390439,
+      "learning_rate": 7.342148760330579e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -446.0,
+      "logps/rejected": -334.0,
+      "loss": 0.288,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.09375,
+      "step": 224
+    },
+    {
+      "epoch": 0.27607361963190186,
+      "grad_norm": 37.9965018970918,
+      "learning_rate": 7.338842975206612e-07,
+      "logits/chosen": -0.10546875,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -336.0,
+      "logps/rejected": -304.0,
+      "loss": 0.261,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.40625,
+      "step": 225
+    },
+    {
+      "epoch": 0.2773006134969325,
+      "grad_norm": 38.80764018954137,
+      "learning_rate": 7.335537190082644e-07,
+      "logits/chosen": -0.119140625,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3311,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -3.890625,
+      "step": 226
+    },
+    {
+      "epoch": 0.2785276073619632,
+      "grad_norm": 42.55177823975192,
+      "learning_rate": 7.332231404958677e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3198,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.40625,
+      "step": 227
+    },
+    {
+      "epoch": 0.27975460122699386,
+      "grad_norm": 47.83061041994659,
+      "learning_rate": 7.32892561983471e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3258,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -4.125,
+      "step": 228
+    },
+    {
+      "epoch": 0.2809815950920245,
+      "grad_norm": 36.471929966124165,
+      "learning_rate": 7.325619834710743e-07,
+      "logits/chosen": -0.07373046875,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2558,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.59375,
+      "step": 229
+    },
+    {
+      "epoch": 0.2822085889570552,
+      "grad_norm": 36.373451110997884,
+      "learning_rate": 7.322314049586777e-07,
+      "logits/chosen": -0.06884765625,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2454,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.28125,
+      "step": 230
+    },
+    {
+      "epoch": 0.28343558282208586,
+      "grad_norm": 44.837599338102365,
+      "learning_rate": 7.31900826446281e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3895,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.625,
+      "step": 231
+    },
+    {
+      "epoch": 0.2846625766871166,
+      "grad_norm": 40.26904672106654,
+      "learning_rate": 7.315702479338843e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3097,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -4.0625,
+      "step": 232
+    },
+    {
+      "epoch": 0.28588957055214725,
+      "grad_norm": 37.854698103189165,
+      "learning_rate": 7.312396694214876e-07,
+      "logits/chosen": -0.10888671875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -288.0,
+      "loss": 0.3055,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.0625,
+      "step": 233
+    },
+    {
+      "epoch": 0.2871165644171779,
+      "grad_norm": 38.018318477115464,
+      "learning_rate": 7.309090909090909e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2977,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.71875,
+      "step": 234
+    },
+    {
+      "epoch": 0.2883435582822086,
+      "grad_norm": 37.47355891059506,
+      "learning_rate": 7.305785123966941e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -320.0,
+      "loss": 0.284,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.0625,
+      "step": 235
+    },
+    {
+      "epoch": 0.28957055214723926,
+      "grad_norm": 53.75756263750908,
+      "learning_rate": 7.302479338842975e-07,
+      "logits/chosen": -0.0238037109375,
+      "logits/rejected": -0.2080078125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -368.0,
+      "loss": 0.4235,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.1875,
+      "step": 236
+    },
+    {
+      "epoch": 0.2907975460122699,
+      "grad_norm": 34.35888091553625,
+      "learning_rate": 7.299173553719008e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2606,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.09375,
+      "step": 237
+    },
+    {
+      "epoch": 0.2920245398773006,
+      "grad_norm": 42.88801905010467,
+      "learning_rate": 7.295867768595041e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3538,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7578125,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -3.90625,
+      "step": 238
+    },
+    {
+      "epoch": 0.29325153374233126,
+      "grad_norm": 42.29179859412283,
+      "learning_rate": 7.292561983471074e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3291,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.15625,
+      "step": 239
+    },
+    {
+      "epoch": 0.294478527607362,
+      "grad_norm": 37.481940316076816,
+      "learning_rate": 7.289256198347107e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -360.0,
+      "loss": 0.2328,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.890625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -4.6875,
+      "step": 240
+    },
+    {
+      "epoch": 0.29570552147239265,
+      "grad_norm": 37.13709897048156,
+      "learning_rate": 7.28595041322314e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2698,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.25,
+      "step": 241
+    },
+    {
+      "epoch": 0.2969325153374233,
+      "grad_norm": 30.985673644793227,
+      "learning_rate": 7.282644628099174e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -318.0,
+      "loss": 0.228,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.84375,
+      "step": 242
+    },
+    {
+      "epoch": 0.298159509202454,
+      "grad_norm": 53.9279211325065,
+      "learning_rate": 7.279338842975206e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -364.0,
+      "loss": 0.3208,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.34375,
+      "step": 243
+    },
+    {
+      "epoch": 0.29938650306748466,
+      "grad_norm": 44.157193824232536,
+      "learning_rate": 7.276033057851239e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -346.0,
+      "loss": 0.4473,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.125,
+      "step": 244
+    },
+    {
+      "epoch": 0.3006134969325153,
+      "grad_norm": 36.965116220324916,
+      "learning_rate": 7.272727272727272e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2506,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.46875,
+      "step": 245
+    },
+    {
+      "epoch": 0.301840490797546,
+      "grad_norm": 35.79905789258026,
+      "learning_rate": 7.269421487603305e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.292,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.078125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.40625,
+      "step": 246
+    },
+    {
+      "epoch": 0.3030674846625767,
+      "grad_norm": 46.030369589419166,
+      "learning_rate": 7.266115702479338e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3237,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.75,
+      "step": 247
+    },
+    {
+      "epoch": 0.3042944785276074,
+      "grad_norm": 38.420936945371935,
+      "learning_rate": 7.262809917355372e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2497,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.28125,
+      "step": 248
+    },
+    {
+      "epoch": 0.30552147239263805,
+      "grad_norm": 42.67532255005307,
+      "learning_rate": 7.259504132231405e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -476.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2862,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.328125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.3125,
+      "step": 249
+    },
+    {
+      "epoch": 0.3067484662576687,
+      "grad_norm": 40.02525959257945,
+      "learning_rate": 7.256198347107438e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -342.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3299,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -4.40625,
+      "step": 250
+    },
+    {
+      "epoch": 0.3079754601226994,
+      "grad_norm": 43.266430285019,
+      "learning_rate": 7.252892561983471e-07,
+      "logits/chosen": -0.041015625,
+      "logits/rejected": -0.2138671875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3567,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.75,
+      "step": 251
+    },
+    {
+      "epoch": 0.30920245398773005,
+      "grad_norm": 37.89906733852128,
+      "learning_rate": 7.249586776859503e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2982,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.7734375,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -3.859375,
+      "step": 252
+    },
+    {
+      "epoch": 0.3104294478527607,
+      "grad_norm": 33.28018552295872,
+      "learning_rate": 7.246280991735536e-07,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2522,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.46875,
+      "step": 253
+    },
+    {
+      "epoch": 0.3116564417177914,
+      "grad_norm": 42.254563825083,
+      "learning_rate": 7.24297520661157e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3681,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.15625,
+      "step": 254
+    },
+    {
+      "epoch": 0.3128834355828221,
+      "grad_norm": 39.895653943663504,
+      "learning_rate": 7.239669421487603e-07,
+      "logits/chosen": -0.1396484375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3481,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.8359375,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.46875,
+      "step": 255
+    },
+    {
+      "epoch": 0.3141104294478528,
+      "grad_norm": 37.11298367671905,
+      "learning_rate": 7.236363636363636e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2374,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.34375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.0625,
+      "step": 256
+    },
+    {
+      "epoch": 0.31533742331288345,
+      "grad_norm": 44.74542880830349,
+      "learning_rate": 7.233057851239669e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -426.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3118,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -3.546875,
+      "step": 257
+    },
+    {
+      "epoch": 0.3165644171779141,
+      "grad_norm": 34.05504946265699,
+      "learning_rate": 7.229752066115702e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2467,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.8984375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -3.859375,
+      "step": 258
+    },
+    {
+      "epoch": 0.3177914110429448,
+      "grad_norm": 35.06646898219457,
+      "learning_rate": 7.226446280991736e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -344.0,
+      "loss": 0.278,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.96875,
+      "step": 259
+    },
+    {
+      "epoch": 0.31901840490797545,
+      "grad_norm": 48.15030024240453,
+      "learning_rate": 7.223140495867769e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3099,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -3.640625,
+      "step": 260
+    },
+    {
+      "epoch": 0.3202453987730061,
+      "grad_norm": 43.183782239877885,
+      "learning_rate": 7.219834710743801e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3207,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.71875,
+      "step": 261
+    },
+    {
+      "epoch": 0.3214723926380368,
+      "grad_norm": 43.81473087786423,
+      "learning_rate": 7.216528925619834e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3077,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -3.484375,
+      "step": 262
+    },
+    {
+      "epoch": 0.3226993865030675,
+      "grad_norm": 45.763039013247074,
+      "learning_rate": 7.213223140495867e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -304.0,
+      "loss": 0.345,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -3.265625,
+      "step": 263
+    },
+    {
+      "epoch": 0.3239263803680982,
+      "grad_norm": 41.54807658858699,
+      "learning_rate": 7.2099173553719e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -288.0,
+      "loss": 0.3193,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -3.234375,
+      "step": 264
+    },
+    {
+      "epoch": 0.32515337423312884,
+      "grad_norm": 43.72770395576322,
+      "learning_rate": 7.206611570247934e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -266.0,
+      "loss": 0.296,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 4.96875,
+      "rewards/rejected": -3.515625,
+      "step": 265
+    },
+    {
+      "epoch": 0.3263803680981595,
+      "grad_norm": 49.68472762312363,
+      "learning_rate": 7.203305785123967e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -316.0,
+      "logps/rejected": -292.0,
+      "loss": 0.41,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 0.68359375,
+      "rewards/margins": 3.875,
+      "rewards/rejected": -3.1875,
+      "step": 266
+    },
+    {
+      "epoch": 0.3276073619631902,
+      "grad_norm": 54.21686609921306,
+      "learning_rate": 7.2e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3765,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.828125,
+      "rewards/margins": 4.8125,
+      "rewards/rejected": -2.984375,
+      "step": 267
+    },
+    {
+      "epoch": 0.32883435582822085,
+      "grad_norm": 36.013473849533085,
+      "learning_rate": 7.196694214876033e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2353,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.328125,
+      "step": 268
+    },
+    {
+      "epoch": 0.3300613496932515,
+      "grad_norm": 37.07985683341205,
+      "learning_rate": 7.193388429752066e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -400.0,
+      "logps/rejected": -292.0,
+      "loss": 0.229,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -3.46875,
+      "step": 269
+    },
+    {
+      "epoch": 0.3312883435582822,
+      "grad_norm": 43.39229519288019,
+      "learning_rate": 7.190082644628098e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -326.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3693,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 4.1875,
+      "rewards/rejected": -2.875,
+      "step": 270
+    },
+    {
+      "epoch": 0.3325153374233129,
+      "grad_norm": 33.45853348793219,
+      "learning_rate": 7.186776859504132e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2511,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.34375,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.234375,
+      "step": 271
+    },
+    {
+      "epoch": 0.3337423312883436,
+      "grad_norm": 36.86544969536763,
+      "learning_rate": 7.183471074380165e-07,
+      "logits/chosen": -0.1025390625,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2449,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.640625,
+      "step": 272
+    },
+    {
+      "epoch": 0.33496932515337424,
+      "grad_norm": 45.60464590137911,
+      "learning_rate": 7.180165289256198e-07,
+      "logits/chosen": -0.05029296875,
+      "logits/rejected": -0.1357421875,
+      "logps/chosen": -334.0,
+      "logps/rejected": -292.0,
+      "loss": 0.4336,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.4375,
+      "step": 273
+    },
+    {
+      "epoch": 0.3361963190184049,
+      "grad_norm": 38.72003382453235,
+      "learning_rate": 7.176859504132231e-07,
+      "logits/chosen": -0.06591796875,
+      "logits/rejected": -0.21875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3223,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.359375,
+      "step": 274
+    },
+    {
+      "epoch": 0.3374233128834356,
+      "grad_norm": 41.46753745108232,
+      "learning_rate": 7.173553719008264e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -316.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3348,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.5,
+      "step": 275
+    },
+    {
+      "epoch": 0.33865030674846625,
+      "grad_norm": 29.942801211334718,
+      "learning_rate": 7.170247933884297e-07,
+      "logits/chosen": -0.056396484375,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2861,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.84375,
+      "step": 276
+    },
+    {
+      "epoch": 0.3398773006134969,
+      "grad_norm": 36.649252691867474,
+      "learning_rate": 7.166942148760331e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2238,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.3125,
+      "step": 277
+    },
+    {
+      "epoch": 0.3411042944785276,
+      "grad_norm": 41.64219712301306,
+      "learning_rate": 7.163636363636364e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2881,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.125,
+      "step": 278
+    },
+    {
+      "epoch": 0.3423312883435583,
+      "grad_norm": 25.028100153022006,
+      "learning_rate": 7.160330578512396e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1819,
+      "rewards/accuracies": 0.9453125,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.875,
+      "step": 279
+    },
+    {
+      "epoch": 0.34355828220858897,
+      "grad_norm": 43.34524582256192,
+      "learning_rate": 7.157024793388429e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3811,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.625,
+      "step": 280
+    },
+    {
+      "epoch": 0.34478527607361964,
+      "grad_norm": 39.57499610045821,
+      "learning_rate": 7.153719008264462e-07,
+      "logits/chosen": -0.12158203125,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3441,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.9375,
+      "step": 281
+    },
+    {
+      "epoch": 0.3460122699386503,
+      "grad_norm": 32.558914994483004,
+      "learning_rate": 7.150413223140495e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -340.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2162,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.0625,
+      "step": 282
+    },
+    {
+      "epoch": 0.347239263803681,
+      "grad_norm": 33.09320412279389,
+      "learning_rate": 7.147107438016529e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2537,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.984375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.5,
+      "step": 283
+    },
+    {
+      "epoch": 0.34846625766871164,
+      "grad_norm": 40.734540729970675,
+      "learning_rate": 7.143801652892562e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -310.0,
+      "loss": 0.272,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.84375,
+      "step": 284
+    },
+    {
+      "epoch": 0.3496932515337423,
+      "grad_norm": 30.9410258264426,
+      "learning_rate": 7.140495867768595e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2413,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.96875,
+      "step": 285
+    },
+    {
+      "epoch": 0.350920245398773,
+      "grad_norm": 37.246332837421065,
+      "learning_rate": 7.137190082644628e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3337,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.09375,
+      "step": 286
+    },
+    {
+      "epoch": 0.3521472392638037,
+      "grad_norm": 40.87355956326826,
+      "learning_rate": 7.133884297520661e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2981,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.68359375,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.9375,
+      "step": 287
+    },
+    {
+      "epoch": 0.35337423312883437,
+      "grad_norm": 37.148615900618886,
+      "learning_rate": 7.130578512396693e-07,
+      "logits/chosen": -0.05908203125,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3085,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -4.59375,
+      "step": 288
+    },
+    {
+      "epoch": 0.35460122699386504,
+      "grad_norm": 29.32351813266847,
+      "learning_rate": 7.127272727272727e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -316.0,
+      "loss": 0.1987,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.34375,
+      "step": 289
+    },
+    {
+      "epoch": 0.3558282208588957,
+      "grad_norm": 34.03478605235303,
+      "learning_rate": 7.12396694214876e-07,
+      "logits/chosen": -0.359375,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2024,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.71875,
+      "step": 290
+    },
+    {
+      "epoch": 0.3570552147239264,
+      "grad_norm": 38.57928856233812,
+      "learning_rate": 7.120661157024793e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3224,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -4.5,
+      "step": 291
+    },
+    {
+      "epoch": 0.35828220858895704,
+      "grad_norm": 36.5133266014805,
+      "learning_rate": 7.117355371900826e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2451,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.03125,
+      "step": 292
+    },
+    {
+      "epoch": 0.3595092024539877,
+      "grad_norm": 31.073871401773395,
+      "learning_rate": 7.114049586776859e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2162,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.25,
+      "step": 293
+    },
+    {
+      "epoch": 0.36073619631901843,
+      "grad_norm": 31.419020907335653,
+      "learning_rate": 7.110743801652892e-07,
+      "logits/chosen": -0.427734375,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -464.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2389,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.3828125,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.53125,
+      "step": 294
+    },
+    {
+      "epoch": 0.3619631901840491,
+      "grad_norm": 33.66575248055817,
+      "learning_rate": 7.107438016528926e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2242,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.03125,
+      "step": 295
+    },
+    {
+      "epoch": 0.36319018404907977,
+      "grad_norm": 40.06405186289625,
+      "learning_rate": 7.104132231404959e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3209,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.1875,
+      "step": 296
+    },
+    {
+      "epoch": 0.36441717791411044,
+      "grad_norm": 30.371449913820346,
+      "learning_rate": 7.100826446280991e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -352.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2354,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.5,
+      "step": 297
+    },
+    {
+      "epoch": 0.3656441717791411,
+      "grad_norm": 40.780831154790015,
+      "learning_rate": 7.097520661157024e-07,
+      "logits/chosen": -0.126953125,
+      "logits/rejected": -0.23046875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2901,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.890625,
+      "step": 298
+    },
+    {
+      "epoch": 0.36687116564417177,
+      "grad_norm": 48.39132901882327,
+      "learning_rate": 7.094214876033057e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3427,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -4.0625,
+      "step": 299
+    },
+    {
+      "epoch": 0.36809815950920244,
+      "grad_norm": 39.25730980280545,
+      "learning_rate": 7.09090909090909e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3287,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.21875,
+      "step": 300
+    },
+    {
+      "epoch": 0.3693251533742331,
+      "grad_norm": 34.87785766178715,
+      "learning_rate": 7.087603305785124e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -448.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2823,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.1875,
+      "step": 301
+    },
+    {
+      "epoch": 0.37055214723926383,
+      "grad_norm": 30.464291800676158,
+      "learning_rate": 7.084297520661157e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2182,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -4.15625,
+      "step": 302
+    },
+    {
+      "epoch": 0.3717791411042945,
+      "grad_norm": 45.78814856158198,
+      "learning_rate": 7.08099173553719e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3512,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.40625,
+      "step": 303
+    },
+    {
+      "epoch": 0.37300613496932516,
+      "grad_norm": 33.26834608933062,
+      "learning_rate": 7.077685950413223e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2448,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.53125,
+      "step": 304
+    },
+    {
+      "epoch": 0.37423312883435583,
+      "grad_norm": 35.609604648743236,
+      "learning_rate": 7.074380165289256e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2744,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.625,
+      "step": 305
+    },
+    {
+      "epoch": 0.3754601226993865,
+      "grad_norm": 43.837358024220855,
+      "learning_rate": 7.071074380165288e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2737,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.5,
+      "step": 306
+    },
+    {
+      "epoch": 0.37668711656441717,
+      "grad_norm": 39.17650253748135,
+      "learning_rate": 7.067768595041322e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3713,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.71875,
+      "step": 307
+    },
+    {
+      "epoch": 0.37791411042944784,
+      "grad_norm": 36.807676017880326,
+      "learning_rate": 7.064462809917355e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3165,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.94140625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -4.0625,
+      "step": 308
+    },
+    {
+      "epoch": 0.3791411042944785,
+      "grad_norm": 32.288551419968854,
+      "learning_rate": 7.061157024793388e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2283,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.6484375,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -4.90625,
+      "step": 309
+    },
+    {
+      "epoch": 0.3803680981595092,
+      "grad_norm": 47.130716517812814,
+      "learning_rate": 7.057851239669421e-07,
+      "logits/chosen": -0.0908203125,
+      "logits/rejected": -0.1884765625,
+      "logps/chosen": -312.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3596,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.15234375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.6875,
+      "step": 310
+    },
+    {
+      "epoch": 0.3815950920245399,
+      "grad_norm": 34.12148114082454,
+      "learning_rate": 7.054545454545454e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2543,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -5.375,
+      "step": 311
+    },
+    {
+      "epoch": 0.38282208588957056,
+      "grad_norm": 44.35564676375722,
+      "learning_rate": 7.051239669421488e-07,
+      "logits/chosen": -0.125,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -362.0,
+      "loss": 0.4044,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -4.5,
+      "step": 312
+    },
+    {
+      "epoch": 0.38404907975460123,
+      "grad_norm": 33.82923751151301,
+      "learning_rate": 7.047933884297521e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2685,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.8828125,
+      "rewards/margins": 7.28125,
+      "rewards/rejected": -5.375,
+      "step": 313
+    },
+    {
+      "epoch": 0.3852760736196319,
+      "grad_norm": 33.17604766084792,
+      "learning_rate": 7.044628099173554e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -456.0,
+      "logps/rejected": -344.0,
+      "loss": 0.1847,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 2.25,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -5.71875,
+      "step": 314
+    },
+    {
+      "epoch": 0.38650306748466257,
+      "grad_norm": 38.90508166705689,
+      "learning_rate": 7.041322314049586e-07,
+      "logits/chosen": -0.06396484375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -286.0,
+      "loss": 0.2412,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.5546875,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.21875,
+      "step": 315
+    },
+    {
+      "epoch": 0.38773006134969323,
+      "grad_norm": 31.950078808944912,
+      "learning_rate": 7.038016528925619e-07,
+      "logits/chosen": -0.06982421875,
+      "logits/rejected": -0.2158203125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2551,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.4375,
+      "step": 316
+    },
+    {
+      "epoch": 0.3889570552147239,
+      "grad_norm": 34.528377745666106,
+      "learning_rate": 7.034710743801652e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2296,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.875,
+      "step": 317
+    },
+    {
+      "epoch": 0.3901840490797546,
+      "grad_norm": 46.9506582652498,
+      "learning_rate": 7.031404958677686e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3503,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.71875,
+      "step": 318
+    },
+    {
+      "epoch": 0.3914110429447853,
+      "grad_norm": 26.867054074267944,
+      "learning_rate": 7.028099173553719e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1701,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 7.4375,
+      "rewards/rejected": -6.125,
+      "step": 319
+    },
+    {
+      "epoch": 0.39263803680981596,
+      "grad_norm": 40.540053142562215,
+      "learning_rate": 7.024793388429752e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -318.0,
+      "logps/rejected": -274.0,
+      "loss": 0.3046,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -5.3125,
+      "step": 320
+    },
+    {
+      "epoch": 0.39386503067484663,
+      "grad_norm": 46.49642012587718,
+      "learning_rate": 7.021487603305785e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -384.0,
+      "loss": 0.2961,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.96875,
+      "step": 321
+    },
+    {
+      "epoch": 0.3950920245398773,
+      "grad_norm": 33.72622281793196,
+      "learning_rate": 7.018181818181818e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2496,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.3125,
+      "step": 322
+    },
+    {
+      "epoch": 0.39631901840490796,
+      "grad_norm": 39.04984997032008,
+      "learning_rate": 7.014876033057851e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3837,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -4.3125,
+      "step": 323
+    },
+    {
+      "epoch": 0.39754601226993863,
+      "grad_norm": 33.57036215026917,
+      "learning_rate": 7.011570247933884e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2714,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.5,
+      "step": 324
+    },
+    {
+      "epoch": 0.3987730061349693,
+      "grad_norm": 45.64995516248102,
+      "learning_rate": 7.008264462809917e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.2177734375,
+      "logps/chosen": -328.0,
+      "logps/rejected": -290.0,
+      "loss": 0.4112,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.6875,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -4.53125,
+      "step": 325
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 34.28943945713478,
+      "learning_rate": 7.00495867768595e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2004,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.09375,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -5.03125,
+      "step": 326
+    },
+    {
+      "epoch": 0.4012269938650307,
+      "grad_norm": 48.50197304580951,
+      "learning_rate": 7.001652892561983e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3949,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.6875,
+      "step": 327
+    },
+    {
+      "epoch": 0.40245398773006136,
+      "grad_norm": 40.72777385232644,
+      "learning_rate": 6.998347107438016e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -272.0,
+      "loss": 0.3146,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -4.3125,
+      "step": 328
+    },
+    {
+      "epoch": 0.403680981595092,
+      "grad_norm": 35.55525281765151,
+      "learning_rate": 6.995041322314049e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -454.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2936,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -4.90625,
+      "step": 329
+    },
+    {
+      "epoch": 0.4049079754601227,
+      "grad_norm": 42.0221302571432,
+      "learning_rate": 6.991735537190083e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2435,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9765625,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -4.71875,
+      "step": 330
+    },
+    {
+      "epoch": 0.40613496932515336,
+      "grad_norm": 32.84700653520424,
+      "learning_rate": 6.988429752066116e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2501,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.9375,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -4.375,
+      "step": 331
+    },
+    {
+      "epoch": 0.40736196319018403,
+      "grad_norm": 27.379927529722597,
+      "learning_rate": 6.985123966942148e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2017,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.46875,
+      "step": 332
+    },
+    {
+      "epoch": 0.4085889570552147,
+      "grad_norm": 36.70659629756495,
+      "learning_rate": 6.981818181818181e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3078,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -3.84375,
+      "step": 333
+    },
+    {
+      "epoch": 0.4098159509202454,
+      "grad_norm": 34.15971280742034,
+      "learning_rate": 6.978512396694214e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -338.0,
+      "loss": 0.273,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -4.5,
+      "step": 334
+    },
+    {
+      "epoch": 0.4110429447852761,
+      "grad_norm": 37.2375068536928,
+      "learning_rate": 6.975206611570247e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2248,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.03125,
+      "step": 335
+    },
+    {
+      "epoch": 0.41226993865030676,
+      "grad_norm": 36.84981287485191,
+      "learning_rate": 6.971900826446281e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2764,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.21875,
+      "step": 336
+    },
+    {
+      "epoch": 0.4134969325153374,
+      "grad_norm": 46.94122451213228,
+      "learning_rate": 6.968595041322314e-07,
+      "logits/chosen": -0.09375,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -344.0,
+      "logps/rejected": -322.0,
+      "loss": 0.389,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.921875,
+      "step": 337
+    },
+    {
+      "epoch": 0.4147239263803681,
+      "grad_norm": 38.247859512270246,
+      "learning_rate": 6.965289256198347e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3295,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.0,
+      "step": 338
+    },
+    {
+      "epoch": 0.41595092024539876,
+      "grad_norm": 40.84741702132196,
+      "learning_rate": 6.96198347107438e-07,
+      "logits/chosen": -0.07470703125,
+      "logits/rejected": -0.2109375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3275,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.15625,
+      "step": 339
+    },
+    {
+      "epoch": 0.4171779141104294,
+      "grad_norm": 43.700613878566976,
+      "learning_rate": 6.958677685950413e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2879,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.09375,
+      "step": 340
+    },
+    {
+      "epoch": 0.41840490797546015,
+      "grad_norm": 31.54531879054049,
+      "learning_rate": 6.955371900826445e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2805,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.65625,
+      "step": 341
+    },
+    {
+      "epoch": 0.4196319018404908,
+      "grad_norm": 39.0902727508844,
+      "learning_rate": 6.952066115702479e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -328.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3322,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.1875,
+      "step": 342
+    },
+    {
+      "epoch": 0.4208588957055215,
+      "grad_norm": 39.32051666516945,
+      "learning_rate": 6.948760330578512e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3545,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.921875,
+      "step": 343
+    },
+    {
+      "epoch": 0.42208588957055215,
+      "grad_norm": 42.12526028461151,
+      "learning_rate": 6.945454545454545e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3584,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.0625,
+      "step": 344
+    },
+    {
+      "epoch": 0.4233128834355828,
+      "grad_norm": 39.02371998906191,
+      "learning_rate": 6.942148760330578e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3288,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.3125,
+      "step": 345
+    },
+    {
+      "epoch": 0.4245398773006135,
+      "grad_norm": 41.19695447421548,
+      "learning_rate": 6.938842975206611e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -332.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3569,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -4.125,
+      "step": 346
+    },
+    {
+      "epoch": 0.42576687116564416,
+      "grad_norm": 28.472619739287644,
+      "learning_rate": 6.935537190082645e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -360.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2332,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.25,
+      "step": 347
+    },
+    {
+      "epoch": 0.4269938650306748,
+      "grad_norm": 33.19419614917591,
+      "learning_rate": 6.932231404958678e-07,
+      "logits/chosen": -0.24609375,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2594,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.91796875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.59375,
+      "step": 348
+    },
+    {
+      "epoch": 0.42822085889570555,
+      "grad_norm": 38.111059722172605,
+      "learning_rate": 6.928925619834711e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3255,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.34375,
+      "step": 349
+    },
+    {
+      "epoch": 0.4294478527607362,
+      "grad_norm": 37.49784569992658,
+      "learning_rate": 6.925619834710743e-07,
+      "logits/chosen": -0.0986328125,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2857,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.28125,
+      "step": 350
+    },
+    {
+      "epoch": 0.4306748466257669,
+      "grad_norm": 42.647163206760126,
+      "learning_rate": 6.922314049586776e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -336.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3505,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.5078125,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -4.78125,
+      "step": 351
+    },
+    {
+      "epoch": 0.43190184049079755,
+      "grad_norm": 36.26120743659289,
+      "learning_rate": 6.919008264462809e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3453,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.03125,
+      "step": 352
+    },
+    {
+      "epoch": 0.4331288343558282,
+      "grad_norm": 30.933441803520967,
+      "learning_rate": 6.915702479338843e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2406,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.96875,
+      "step": 353
+    },
+    {
+      "epoch": 0.4343558282208589,
+      "grad_norm": 34.874499707263155,
+      "learning_rate": 6.912396694214876e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2963,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.03125,
+      "step": 354
+    },
+    {
+      "epoch": 0.43558282208588955,
+      "grad_norm": 39.848840136754845,
+      "learning_rate": 6.909090909090909e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -340.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3581,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.53125,
+      "step": 355
+    },
+    {
+      "epoch": 0.4368098159509202,
+      "grad_norm": 44.03875893573232,
+      "learning_rate": 6.905785123966942e-07,
+      "logits/chosen": -0.0966796875,
+      "logits/rejected": -0.240234375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3406,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.03125,
+      "step": 356
+    },
+    {
+      "epoch": 0.43803680981595094,
+      "grad_norm": 36.68968799262988,
+      "learning_rate": 6.902479338842975e-07,
+      "logits/chosen": -0.3515625,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2926,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -4.9375,
+      "step": 357
+    },
+    {
+      "epoch": 0.4392638036809816,
+      "grad_norm": 36.204158508103596,
+      "learning_rate": 6.899173553719008e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2673,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.40625,
+      "step": 358
+    },
+    {
+      "epoch": 0.4404907975460123,
+      "grad_norm": 40.16138545949422,
+      "learning_rate": 6.89586776859504e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.953125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.09375,
+      "step": 359
+    },
+    {
+      "epoch": 0.44171779141104295,
+      "grad_norm": 35.21609251200673,
+      "learning_rate": 6.892561983471074e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2614,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.9375,
+      "step": 360
+    },
+    {
+      "epoch": 0.4429447852760736,
+      "grad_norm": 33.95190421090596,
+      "learning_rate": 6.889256198347107e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2599,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -4.8125,
+      "step": 361
+    },
+    {
+      "epoch": 0.4441717791411043,
+      "grad_norm": 32.41395226001234,
+      "learning_rate": 6.88595041322314e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2625,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -4.75,
+      "step": 362
+    },
+    {
+      "epoch": 0.44539877300613495,
+      "grad_norm": 31.06408733284358,
+      "learning_rate": 6.882644628099173e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2362,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.375,
+      "step": 363
+    },
+    {
+      "epoch": 0.4466257668711656,
+      "grad_norm": 40.778000027843554,
+      "learning_rate": 6.879338842975206e-07,
+      "logits/chosen": -0.07470703125,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2885,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.28125,
+      "step": 364
+    },
+    {
+      "epoch": 0.44785276073619634,
+      "grad_norm": 30.8545654380382,
+      "learning_rate": 6.87603305785124e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -298.0,
+      "loss": 0.2419,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.53125,
+      "step": 365
+    },
+    {
+      "epoch": 0.449079754601227,
+      "grad_norm": 35.25923514975111,
+      "learning_rate": 6.872727272727273e-07,
+      "logits/chosen": -0.11376953125,
+      "logits/rejected": -0.21484375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3324,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.71875,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.703125,
+      "step": 366
+    },
+    {
+      "epoch": 0.4503067484662577,
+      "grad_norm": 29.75727838549342,
+      "learning_rate": 6.869421487603306e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2265,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.25,
+      "step": 367
+    },
+    {
+      "epoch": 0.45153374233128835,
+      "grad_norm": 41.34860966923763,
+      "learning_rate": 6.866115702479338e-07,
+      "logits/chosen": -0.060546875,
+      "logits/rejected": -0.11328125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3251,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.1875,
+      "step": 368
+    },
+    {
+      "epoch": 0.452760736196319,
+      "grad_norm": 39.802701196225414,
+      "learning_rate": 6.862809917355371e-07,
+      "logits/chosen": -0.08349609375,
+      "logits/rejected": -0.19140625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3103,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -4.375,
+      "step": 369
+    },
+    {
+      "epoch": 0.4539877300613497,
+      "grad_norm": 35.44626532768664,
+      "learning_rate": 6.859504132231404e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3001,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.0,
+      "step": 370
+    },
+    {
+      "epoch": 0.45521472392638035,
+      "grad_norm": 34.081640470685045,
+      "learning_rate": 6.856198347107438e-07,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -458.0,
+      "logps/rejected": -356.0,
+      "loss": 0.268,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.734375,
+      "step": 371
+    },
+    {
+      "epoch": 0.456441717791411,
+      "grad_norm": 32.10146522336104,
+      "learning_rate": 6.852892561983471e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2227,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.15625,
+      "step": 372
+    },
+    {
+      "epoch": 0.45766871165644174,
+      "grad_norm": 38.40212256087746,
+      "learning_rate": 6.849586776859504e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -360.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3491,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.88671875,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -4.40625,
+      "step": 373
+    },
+    {
+      "epoch": 0.4588957055214724,
+      "grad_norm": 33.89078532536025,
+      "learning_rate": 6.846280991735537e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2439,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.25,
+      "step": 374
+    },
+    {
+      "epoch": 0.4601226993865031,
+      "grad_norm": 38.36603856770271,
+      "learning_rate": 6.84297520661157e-07,
+      "logits/chosen": -0.24609375,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2743,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.78125,
+      "step": 375
+    },
+    {
+      "epoch": 0.46134969325153374,
+      "grad_norm": 32.73897251445023,
+      "learning_rate": 6.839669421487603e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -342.0,
+      "loss": 0.231,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -3.9375,
+      "step": 376
+    },
+    {
+      "epoch": 0.4625766871165644,
+      "grad_norm": 36.44881761827906,
+      "learning_rate": 6.836363636363636e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -350.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2817,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.8125,
+      "step": 377
+    },
+    {
+      "epoch": 0.4638036809815951,
+      "grad_norm": 36.92866834323212,
+      "learning_rate": 6.833057851239669e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3001,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.5,
+      "step": 378
+    },
+    {
+      "epoch": 0.46503067484662575,
+      "grad_norm": 32.73144242036404,
+      "learning_rate": 6.829752066115702e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2471,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7578125,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.0625,
+      "step": 379
+    },
+    {
+      "epoch": 0.4662576687116564,
+      "grad_norm": 34.6763093414022,
+      "learning_rate": 6.826446280991735e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2659,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -4.96875,
+      "step": 380
+    },
+    {
+      "epoch": 0.46748466257668714,
+      "grad_norm": 34.98446202126656,
+      "learning_rate": 6.823140495867768e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3193,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.59375,
+      "step": 381
+    },
+    {
+      "epoch": 0.4687116564417178,
+      "grad_norm": 33.068661800772716,
+      "learning_rate": 6.819834710743801e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.40625,
+      "step": 382
+    },
+    {
+      "epoch": 0.4699386503067485,
+      "grad_norm": 37.187464031340724,
+      "learning_rate": 6.816528925619835e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -378.0,
+      "loss": 0.33,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.625,
+      "step": 383
+    },
+    {
+      "epoch": 0.47116564417177914,
+      "grad_norm": 35.133860722289874,
+      "learning_rate": 6.813223140495868e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2466,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -6.5,
+      "step": 384
+    },
+    {
+      "epoch": 0.4723926380368098,
+      "grad_norm": 40.48440514462701,
+      "learning_rate": 6.809917355371901e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3252,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.2578125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -5.78125,
+      "step": 385
+    },
+    {
+      "epoch": 0.4736196319018405,
+      "grad_norm": 40.628340663991764,
+      "learning_rate": 6.806611570247933e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3025,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -0.359375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -6.0625,
+      "step": 386
+    },
+    {
+      "epoch": 0.47484662576687114,
+      "grad_norm": 51.72741620856106,
+      "learning_rate": 6.803305785123966e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.90625,
+      "step": 387
+    },
+    {
+      "epoch": 0.47607361963190187,
+      "grad_norm": 43.50814776279542,
+      "learning_rate": 6.8e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3977,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.05029296875,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -5.09375,
+      "step": 388
+    },
+    {
+      "epoch": 0.47730061349693254,
+      "grad_norm": 44.9083628878753,
+      "learning_rate": 6.796694214876033e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3913,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.44140625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.625,
+      "step": 389
+    },
+    {
+      "epoch": 0.4785276073619632,
+      "grad_norm": 42.893730022875914,
+      "learning_rate": 6.793388429752066e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3759,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": -0.263671875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -6.0,
+      "step": 390
+    },
+    {
+      "epoch": 0.47975460122699387,
+      "grad_norm": 46.56902163562806,
+      "learning_rate": 6.790082644628099e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4122,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.3984375,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.5625,
+      "step": 391
+    },
+    {
+      "epoch": 0.48098159509202454,
+      "grad_norm": 32.46380915314015,
+      "learning_rate": 6.786776859504132e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2695,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.65625,
+      "step": 392
+    },
+    {
+      "epoch": 0.4822085889570552,
+      "grad_norm": 36.411647842869584,
+      "learning_rate": 6.783471074380165e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2936,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.66015625,
+      "rewards/margins": 6.125,
+      "rewards/rejected": -5.46875,
+      "step": 393
+    },
+    {
+      "epoch": 0.4834355828220859,
+      "grad_norm": 36.43693693488016,
+      "learning_rate": 6.780165289256199e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2504,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.7265625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.34375,
+      "step": 394
+    },
+    {
+      "epoch": 0.48466257668711654,
+      "grad_norm": 36.92302249692751,
+      "learning_rate": 6.776859504132231e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3275,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.96875,
+      "step": 395
+    },
+    {
+      "epoch": 0.48588957055214727,
+      "grad_norm": 31.266490118113346,
+      "learning_rate": 6.773553719008264e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -338.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2215,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.462890625,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.6875,
+      "step": 396
+    },
+    {
+      "epoch": 0.48711656441717793,
+      "grad_norm": 34.937182799458625,
+      "learning_rate": 6.770247933884297e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2752,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.8125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.65625,
+      "step": 397
+    },
+    {
+      "epoch": 0.4883435582822086,
+      "grad_norm": 38.419527775162116,
+      "learning_rate": 6.76694214876033e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3231,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.0625,
+      "step": 398
+    },
+    {
+      "epoch": 0.48957055214723927,
+      "grad_norm": 35.54458075122045,
+      "learning_rate": 6.763636363636363e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -330.0,
+      "loss": 0.306,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.9375,
+      "step": 399
+    },
+    {
+      "epoch": 0.49079754601226994,
+      "grad_norm": 47.751486198572856,
+      "learning_rate": 6.760330578512397e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3896,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.90234375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.46875,
+      "step": 400
+    },
+    {
+      "epoch": 0.4920245398773006,
+      "grad_norm": 31.519656830052252,
+      "learning_rate": 6.75702479338843e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -444.0,
+      "logps/rejected": -372.0,
+      "loss": 0.2337,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.34375,
+      "step": 401
+    },
+    {
+      "epoch": 0.49325153374233127,
+      "grad_norm": 36.11809592072752,
+      "learning_rate": 6.753719008264463e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3095,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.78125,
+      "step": 402
+    },
+    {
+      "epoch": 0.49447852760736194,
+      "grad_norm": 38.21909292164752,
+      "learning_rate": 6.750413223140496e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3169,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.875,
+      "step": 403
+    },
+    {
+      "epoch": 0.49570552147239266,
+      "grad_norm": 32.22630836566114,
+      "learning_rate": 6.747107438016528e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2283,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.0,
+      "step": 404
+    },
+    {
+      "epoch": 0.49693251533742333,
+      "grad_norm": 37.593793152840426,
+      "learning_rate": 6.743801652892561e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3305,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.65625,
+      "step": 405
+    },
+    {
+      "epoch": 0.498159509202454,
+      "grad_norm": 41.53333095908596,
+      "learning_rate": 6.740495867768595e-07,
+      "logits/chosen": -0.115234375,
+      "logits/rejected": -0.2109375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3932,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.3125,
+      "step": 406
+    },
+    {
+      "epoch": 0.49938650306748467,
+      "grad_norm": 42.42876322706536,
+      "learning_rate": 6.737190082644628e-07,
+      "logits/chosen": -0.0986328125,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3513,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.140625,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -4.3125,
+      "step": 407
+    },
+    {
+      "epoch": 0.5006134969325153,
+      "grad_norm": 48.89324534838805,
+      "learning_rate": 6.733884297520661e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -348.0,
+      "loss": 0.4164,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 4.78125,
+      "rewards/rejected": -3.984375,
+      "step": 408
+    },
+    {
+      "epoch": 0.501840490797546,
+      "grad_norm": 45.10686444580896,
+      "learning_rate": 6.730578512396694e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3079,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.46875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -4.75,
+      "step": 409
+    },
+    {
+      "epoch": 0.5030674846625767,
+      "grad_norm": 38.62243714595701,
+      "learning_rate": 6.727272727272727e-07,
+      "logits/chosen": -0.119140625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2966,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.3125,
+      "step": 410
+    },
+    {
+      "epoch": 0.5042944785276073,
+      "grad_norm": 37.41136253841519,
+      "learning_rate": 6.72396694214876e-07,
+      "logits/chosen": -0.11083984375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -306.0,
+      "loss": 0.284,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.76953125,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.9375,
+      "step": 411
+    },
+    {
+      "epoch": 0.505521472392638,
+      "grad_norm": 47.69531455944101,
+      "learning_rate": 6.720661157024794e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3619,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -4.28125,
+      "step": 412
+    },
+    {
+      "epoch": 0.5067484662576687,
+      "grad_norm": 39.571364824526874,
+      "learning_rate": 6.717355371900826e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3069,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.59375,
+      "step": 413
+    },
+    {
+      "epoch": 0.5079754601226993,
+      "grad_norm": 34.32054957685153,
+      "learning_rate": 6.714049586776859e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2497,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.8125,
+      "step": 414
+    },
+    {
+      "epoch": 0.50920245398773,
+      "grad_norm": 35.2489180238823,
+      "learning_rate": 6.710743801652892e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2788,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.84375,
+      "step": 415
+    },
+    {
+      "epoch": 0.5104294478527608,
+      "grad_norm": 28.41248357906051,
+      "learning_rate": 6.707438016528925e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -310.0,
+      "loss": 0.186,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 0.8515625,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.75,
+      "step": 416
+    },
+    {
+      "epoch": 0.5116564417177915,
+      "grad_norm": 35.52909474934163,
+      "learning_rate": 6.704132231404958e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -434.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2995,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -5.53125,
+      "step": 417
+    },
+    {
+      "epoch": 0.5128834355828221,
+      "grad_norm": 36.29519060781132,
+      "learning_rate": 6.700826446280992e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -292.0,
+      "loss": 0.2828,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.09375,
+      "step": 418
+    },
+    {
+      "epoch": 0.5141104294478528,
+      "grad_norm": 31.87680208130441,
+      "learning_rate": 6.697520661157025e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2486,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.59375,
+      "step": 419
+    },
+    {
+      "epoch": 0.5153374233128835,
+      "grad_norm": 40.881697279170936,
+      "learning_rate": 6.694214876033058e-07,
+      "logits/chosen": -0.1162109375,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -344.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3244,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.75,
+      "step": 420
+    },
+    {
+      "epoch": 0.5165644171779141,
+      "grad_norm": 34.11082321167536,
+      "learning_rate": 6.69090909090909e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -312.0,
+      "loss": 0.339,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -4.8125,
+      "step": 421
+    },
+    {
+      "epoch": 0.5177914110429448,
+      "grad_norm": 35.18821029472038,
+      "learning_rate": 6.687603305785123e-07,
+      "logits/chosen": -0.0810546875,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3166,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.34375,
+      "step": 422
+    },
+    {
+      "epoch": 0.5190184049079755,
+      "grad_norm": 41.06855790218669,
+      "learning_rate": 6.684297520661156e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3606,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.96875,
+      "step": 423
+    },
+    {
+      "epoch": 0.5202453987730061,
+      "grad_norm": 36.724024878608596,
+      "learning_rate": 6.68099173553719e-07,
+      "logits/chosen": -0.123046875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3438,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.5625,
+      "step": 424
+    },
+    {
+      "epoch": 0.5214723926380368,
+      "grad_norm": 41.01181618114964,
+      "learning_rate": 6.677685950413223e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -342.0,
+      "loss": 0.4073,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.96875,
+      "step": 425
+    },
+    {
+      "epoch": 0.5226993865030675,
+      "grad_norm": 44.64404869891064,
+      "learning_rate": 6.674380165289256e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3035,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.46875,
+      "step": 426
+    },
+    {
+      "epoch": 0.5239263803680981,
+      "grad_norm": 91.8646628453638,
+      "learning_rate": 6.671074380165289e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -328.0,
+      "loss": 0.299,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.09375,
+      "step": 427
+    },
+    {
+      "epoch": 0.5251533742331288,
+      "grad_norm": 36.99313104379447,
+      "learning_rate": 6.667768595041322e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3507,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.66015625,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -5.21875,
+      "step": 428
+    },
+    {
+      "epoch": 0.5263803680981595,
+      "grad_norm": 33.0190569729192,
+      "learning_rate": 6.664462809917356e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -360.0,
+      "logps/rejected": -366.0,
+      "loss": 0.246,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.443359375,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -6.46875,
+      "step": 429
+    },
+    {
+      "epoch": 0.5276073619631901,
+      "grad_norm": 30.408983999620382,
+      "learning_rate": 6.661157024793388e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2248,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.9921875,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.84375,
+      "step": 430
+    },
+    {
+      "epoch": 0.5288343558282208,
+      "grad_norm": 47.87075432548269,
+      "learning_rate": 6.657851239669421e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -336.0,
+      "loss": 0.4545,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.45703125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -5.25,
+      "step": 431
+    },
+    {
+      "epoch": 0.5300613496932516,
+      "grad_norm": 43.65339869428465,
+      "learning_rate": 6.654545454545454e-07,
+      "logits/chosen": -0.0927734375,
+      "logits/rejected": -0.1435546875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3724,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.875,
+      "step": 432
+    },
+    {
+      "epoch": 0.5312883435582823,
+      "grad_norm": 45.133784704459096,
+      "learning_rate": 6.651239669421487e-07,
+      "logits/chosen": -0.078125,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3416,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.39453125,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.34375,
+      "step": 433
+    },
+    {
+      "epoch": 0.5325153374233129,
+      "grad_norm": 40.47930492236747,
+      "learning_rate": 6.64793388429752e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -346.0,
+      "loss": 0.254,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.53125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.25,
+      "step": 434
+    },
+    {
+      "epoch": 0.5337423312883436,
+      "grad_norm": 41.262979059220854,
+      "learning_rate": 6.644628099173554e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3119,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.40625,
+      "step": 435
+    },
+    {
+      "epoch": 0.5349693251533743,
+      "grad_norm": 38.67538014065472,
+      "learning_rate": 6.641322314049587e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2467,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 7.15625,
+      "rewards/rejected": -5.46875,
+      "step": 436
+    },
+    {
+      "epoch": 0.5361963190184049,
+      "grad_norm": 31.32861125197491,
+      "learning_rate": 6.63801652892562e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -338.0,
+      "loss": 0.248,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 6.65625,
+      "rewards/rejected": -5.5625,
+      "step": 437
+    },
+    {
+      "epoch": 0.5374233128834356,
+      "grad_norm": 42.810824724719026,
+      "learning_rate": 6.634710743801653e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.2392578125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3969,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.375,
+      "step": 438
+    },
+    {
+      "epoch": 0.5386503067484663,
+      "grad_norm": 36.79651022084742,
+      "learning_rate": 6.631404958677685e-07,
+      "logits/chosen": -0.083984375,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.1875,
+      "step": 439
+    },
+    {
+      "epoch": 0.5398773006134969,
+      "grad_norm": 49.84881532208057,
+      "learning_rate": 6.628099173553718e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.2333984375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3608,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.384765625,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -4.6875,
+      "step": 440
+    },
+    {
+      "epoch": 0.5411042944785276,
+      "grad_norm": 37.976634967176224,
+      "learning_rate": 6.624793388429752e-07,
+      "logits/chosen": -0.06640625,
+      "logits/rejected": -0.1826171875,
+      "logps/chosen": -332.0,
+      "logps/rejected": -300.0,
+      "loss": 0.328,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.75,
+      "step": 441
+    },
+    {
+      "epoch": 0.5423312883435583,
+      "grad_norm": 37.80188895923485,
+      "learning_rate": 6.621487603305785e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3304,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.8125,
+      "step": 442
+    },
+    {
+      "epoch": 0.5435582822085889,
+      "grad_norm": 40.24381216440989,
+      "learning_rate": 6.618181818181818e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3159,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.4140625,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.59375,
+      "step": 443
+    },
+    {
+      "epoch": 0.5447852760736196,
+      "grad_norm": 37.922945446281,
+      "learning_rate": 6.614876033057851e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -446.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2446,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.09375,
+      "rewards/margins": 7.25,
+      "rewards/rejected": -5.15625,
+      "step": 444
+    },
+    {
+      "epoch": 0.5460122699386503,
+      "grad_norm": 38.45325276499507,
+      "learning_rate": 6.611570247933884e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -300.0,
+      "loss": 0.389,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.125,
+      "step": 445
+    },
+    {
+      "epoch": 0.5472392638036809,
+      "grad_norm": 33.02180393783636,
+      "learning_rate": 6.608264462809917e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2348,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.46875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.25,
+      "step": 446
+    },
+    {
+      "epoch": 0.5484662576687117,
+      "grad_norm": 31.670625439038492,
+      "learning_rate": 6.604958677685951e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2551,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 7.15625,
+      "rewards/rejected": -5.28125,
+      "step": 447
+    },
+    {
+      "epoch": 0.5496932515337424,
+      "grad_norm": 37.605046960174505,
+      "learning_rate": 6.601652892561983e-07,
+      "logits/chosen": -0.10888671875,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3209,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.90234375,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -4.25,
+      "step": 448
+    },
+    {
+      "epoch": 0.550920245398773,
+      "grad_norm": 34.080535896094865,
+      "learning_rate": 6.598347107438016e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2712,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -4.53125,
+      "step": 449
+    },
+    {
+      "epoch": 0.5521472392638037,
+      "grad_norm": 35.069217343013015,
+      "learning_rate": 6.595041322314049e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -288.0,
+      "loss": 0.285,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.625,
+      "step": 450
+    },
+    {
+      "epoch": 0.5533742331288344,
+      "grad_norm": 33.87966228532685,
+      "learning_rate": 6.591735537190082e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -340.0,
+      "logps/rejected": -280.0,
+      "loss": 0.3043,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -4.5,
+      "step": 451
+    },
+    {
+      "epoch": 0.554601226993865,
+      "grad_norm": 35.08833886262592,
+      "learning_rate": 6.588429752066115e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -454.0,
+      "logps/rejected": -328.0,
+      "loss": 0.258,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.421875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.21875,
+      "step": 452
+    },
+    {
+      "epoch": 0.5558282208588957,
+      "grad_norm": 30.509116856340853,
+      "learning_rate": 6.585123966942149e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -366.0,
+      "loss": 0.214,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 2.578125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -4.78125,
+      "step": 453
+    },
+    {
+      "epoch": 0.5570552147239264,
+      "grad_norm": 36.06945734305074,
+      "learning_rate": 6.581818181818182e-07,
+      "logits/chosen": -0.11572265625,
+      "logits/rejected": -0.19921875,
+      "logps/chosen": -338.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2945,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.859375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.625,
+      "step": 454
+    },
+    {
+      "epoch": 0.558282208588957,
+      "grad_norm": 35.814869576892626,
+      "learning_rate": 6.578512396694215e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2399,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.9609375,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -4.8125,
+      "step": 455
+    },
+    {
+      "epoch": 0.5595092024539877,
+      "grad_norm": 38.450506141939925,
+      "learning_rate": 6.575206611570248e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.2236328125,
+      "logps/chosen": -312.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2952,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.125,
+      "step": 456
+    },
+    {
+      "epoch": 0.5607361963190184,
+      "grad_norm": 37.93391723235656,
+      "learning_rate": 6.57190082644628e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3025,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.5625,
+      "step": 457
+    },
+    {
+      "epoch": 0.561963190184049,
+      "grad_norm": 33.44669035850576,
+      "learning_rate": 6.568595041322313e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2297,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.125,
+      "step": 458
+    },
+    {
+      "epoch": 0.5631901840490797,
+      "grad_norm": 34.06851635286411,
+      "learning_rate": 6.565289256198347e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2786,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.0625,
+      "step": 459
+    },
+    {
+      "epoch": 0.5644171779141104,
+      "grad_norm": 34.68751260016757,
+      "learning_rate": 6.56198347107438e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -330.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.76171875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -6.03125,
+      "step": 460
+    },
+    {
+      "epoch": 0.5656441717791411,
+      "grad_norm": 38.71078681224242,
+      "learning_rate": 6.558677685950413e-07,
+      "logits/chosen": -0.3515625,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2741,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.34375,
+      "step": 461
+    },
+    {
+      "epoch": 0.5668711656441717,
+      "grad_norm": 39.327400074833584,
+      "learning_rate": 6.555371900826446e-07,
+      "logits/chosen": -0.11279296875,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3248,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.875,
+      "step": 462
+    },
+    {
+      "epoch": 0.5680981595092025,
+      "grad_norm": 37.637164385131356,
+      "learning_rate": 6.552066115702479e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -360.0,
+      "loss": 0.2848,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.98046875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.46875,
+      "step": 463
+    },
+    {
+      "epoch": 0.5693251533742332,
+      "grad_norm": 41.25059009443085,
+      "learning_rate": 6.548760330578512e-07,
+      "logits/chosen": -0.09619140625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2951,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.34375,
+      "step": 464
+    },
+    {
+      "epoch": 0.5705521472392638,
+      "grad_norm": 27.065991136479923,
+      "learning_rate": 6.545454545454546e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -346.0,
+      "loss": 0.1759,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -6.0625,
+      "step": 465
+    },
+    {
+      "epoch": 0.5717791411042945,
+      "grad_norm": 43.40449571309543,
+      "learning_rate": 6.542148760330578e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2899,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.9375,
+      "step": 466
+    },
+    {
+      "epoch": 0.5730061349693252,
+      "grad_norm": 40.392254031751044,
+      "learning_rate": 6.538842975206611e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3087,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.28125,
+      "step": 467
+    },
+    {
+      "epoch": 0.5742331288343558,
+      "grad_norm": 41.32661857185299,
+      "learning_rate": 6.535537190082644e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3022,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.78125,
+      "step": 468
+    },
+    {
+      "epoch": 0.5754601226993865,
+      "grad_norm": 29.07874137818143,
+      "learning_rate": 6.532231404958677e-07,
+      "logits/chosen": -0.12451171875,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1838,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.6875,
+      "step": 469
+    },
+    {
+      "epoch": 0.5766871165644172,
+      "grad_norm": 35.40917214666752,
+      "learning_rate": 6.52892561983471e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.2353515625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -324.0,
+      "loss": 0.301,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.34375,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -5.40625,
+      "step": 470
+    },
+    {
+      "epoch": 0.5779141104294478,
+      "grad_norm": 38.07545863951338,
+      "learning_rate": 6.525619834710744e-07,
+      "logits/chosen": -0.03369140625,
+      "logits/rejected": -0.203125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3263,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -5.125,
+      "step": 471
+    },
+    {
+      "epoch": 0.5791411042944785,
+      "grad_norm": 40.21918996431336,
+      "learning_rate": 6.522314049586777e-07,
+      "logits/chosen": -0.08447265625,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3173,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -5.15625,
+      "step": 472
+    },
+    {
+      "epoch": 0.5803680981595092,
+      "grad_norm": 37.03158250739783,
+      "learning_rate": 6.51900826446281e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -378.0,
+      "loss": 0.3328,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.90625,
+      "step": 473
+    },
+    {
+      "epoch": 0.5815950920245399,
+      "grad_norm": 38.12266289452875,
+      "learning_rate": 6.515702479338843e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -430.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2648,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -5.21875,
+      "step": 474
+    },
+    {
+      "epoch": 0.5828220858895705,
+      "grad_norm": 31.08159067757818,
+      "learning_rate": 6.512396694214875e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2361,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 7.59375,
+      "rewards/rejected": -5.59375,
+      "step": 475
+    },
+    {
+      "epoch": 0.5840490797546012,
+      "grad_norm": 31.0829148956811,
+      "learning_rate": 6.509090909090908e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2813,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.185546875,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -5.6875,
+      "step": 476
+    },
+    {
+      "epoch": 0.5852760736196319,
+      "grad_norm": 36.7408304979899,
+      "learning_rate": 6.505785123966942e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -354.0,
+      "loss": 0.257,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.75,
+      "step": 477
+    },
+    {
+      "epoch": 0.5865030674846625,
+      "grad_norm": 34.872893059723594,
+      "learning_rate": 6.502479338842975e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.84375,
+      "step": 478
+    },
+    {
+      "epoch": 0.5877300613496933,
+      "grad_norm": 35.26754811760802,
+      "learning_rate": 6.499173553719008e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2784,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.240234375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.96875,
+      "step": 479
+    },
+    {
+      "epoch": 0.588957055214724,
+      "grad_norm": 33.02148763705059,
+      "learning_rate": 6.495867768595041e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2364,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.4765625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -6.125,
+      "step": 480
+    },
+    {
+      "epoch": 0.5901840490797546,
+      "grad_norm": 39.97994667673503,
+      "learning_rate": 6.492561983471074e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -378.0,
+      "loss": 0.3732,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.84765625,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.46875,
+      "step": 481
+    },
+    {
+      "epoch": 0.5914110429447853,
+      "grad_norm": 33.72401411710678,
+      "learning_rate": 6.489256198347108e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2739,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.4375,
+      "step": 482
+    },
+    {
+      "epoch": 0.592638036809816,
+      "grad_norm": 34.19291475160813,
+      "learning_rate": 6.485950413223141e-07,
+      "logits/chosen": -0.0703125,
+      "logits/rejected": -0.2138671875,
+      "logps/chosen": -346.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2917,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.5625,
+      "step": 483
+    },
+    {
+      "epoch": 0.5938650306748466,
+      "grad_norm": 35.398942571150606,
+      "learning_rate": 6.482644628099173e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3155,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.1875,
+      "rewards/rejected": -4.875,
+      "step": 484
+    },
+    {
+      "epoch": 0.5950920245398773,
+      "grad_norm": 30.26872146503766,
+      "learning_rate": 6.479338842975206e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -306.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.625,
+      "step": 485
+    },
+    {
+      "epoch": 0.596319018404908,
+      "grad_norm": 35.38748335338782,
+      "learning_rate": 6.476033057851239e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2658,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -4.96875,
+      "step": 486
+    },
+    {
+      "epoch": 0.5975460122699386,
+      "grad_norm": 30.285435869887902,
+      "learning_rate": 6.472727272727272e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -350.0,
+      "loss": 0.233,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.5625,
+      "step": 487
+    },
+    {
+      "epoch": 0.5987730061349693,
+      "grad_norm": 37.89908939404138,
+      "learning_rate": 6.469421487603306e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2436,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.0625,
+      "step": 488
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 30.515429632482274,
+      "learning_rate": 6.466115702479339e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -326.0,
+      "loss": 0.272,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 7.3125,
+      "rewards/rejected": -5.5,
+      "step": 489
+    },
+    {
+      "epoch": 0.6012269938650306,
+      "grad_norm": 40.37971138033072,
+      "learning_rate": 6.462809917355372e-07,
+      "logits/chosen": -0.09765625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3004,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -5.1875,
+      "step": 490
+    },
+    {
+      "epoch": 0.6024539877300613,
+      "grad_norm": 34.45143017996783,
+      "learning_rate": 6.459504132231405e-07,
+      "logits/chosen": -0.08203125,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2549,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.4375,
+      "step": 491
+    },
+    {
+      "epoch": 0.603680981595092,
+      "grad_norm": 36.58265725949472,
+      "learning_rate": 6.456198347107438e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -378.0,
+      "loss": 0.2737,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.9375,
+      "step": 492
+    },
+    {
+      "epoch": 0.6049079754601226,
+      "grad_norm": 42.46481332966053,
+      "learning_rate": 6.45289256198347e-07,
+      "logits/chosen": -0.024169921875,
+      "logits/rejected": -0.134765625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3917,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -4.25,
+      "step": 493
+    },
+    {
+      "epoch": 0.6061349693251534,
+      "grad_norm": 33.59404853921049,
+      "learning_rate": 6.449586776859504e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2643,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.09375,
+      "step": 494
+    },
+    {
+      "epoch": 0.6073619631901841,
+      "grad_norm": 37.484469870256234,
+      "learning_rate": 6.446280991735537e-07,
+      "logits/chosen": -0.1220703125,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3066,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.21875,
+      "step": 495
+    },
+    {
+      "epoch": 0.6085889570552148,
+      "grad_norm": 40.63137683640751,
+      "learning_rate": 6.44297520661157e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3177,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -5.28125,
+      "step": 496
+    },
+    {
+      "epoch": 0.6098159509202454,
+      "grad_norm": 41.064298986990785,
+      "learning_rate": 6.439669421487603e-07,
+      "logits/chosen": -0.12158203125,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3865,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.85546875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.5,
+      "step": 497
+    },
+    {
+      "epoch": 0.6110429447852761,
+      "grad_norm": 34.6148352569408,
+      "learning_rate": 6.436363636363636e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -358.0,
+      "loss": 0.2817,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.65625,
+      "step": 498
+    },
+    {
+      "epoch": 0.6122699386503068,
+      "grad_norm": 36.60546803390198,
+      "learning_rate": 6.433057851239669e-07,
+      "logits/chosen": -0.115234375,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2842,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.125,
+      "step": 499
+    },
+    {
+      "epoch": 0.6134969325153374,
+      "grad_norm": 33.21675862161022,
+      "learning_rate": 6.429752066115703e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2718,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.21875,
+      "step": 500
+    },
+    {
+      "epoch": 0.6147239263803681,
+      "grad_norm": 32.544310987499145,
+      "learning_rate": 6.426446280991735e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -344.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.625,
+      "step": 501
+    },
+    {
+      "epoch": 0.6159509202453988,
+      "grad_norm": 32.09170254227315,
+      "learning_rate": 6.423140495867768e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2639,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 7.46875,
+      "rewards/rejected": -5.96875,
+      "step": 502
+    },
+    {
+      "epoch": 0.6171779141104294,
+      "grad_norm": 35.53103605940553,
+      "learning_rate": 6.419834710743801e-07,
+      "logits/chosen": 0.01153564453125,
+      "logits/rejected": -0.1904296875,
+      "logps/chosen": -326.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2795,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.25,
+      "step": 503
+    },
+    {
+      "epoch": 0.6184049079754601,
+      "grad_norm": 30.299429390734634,
+      "learning_rate": 6.416528925619834e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2658,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -5.40625,
+      "step": 504
+    },
+    {
+      "epoch": 0.6196319018404908,
+      "grad_norm": 44.689012762744575,
+      "learning_rate": 6.413223140495867e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -394.0,
+      "loss": 0.3776,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.369140625,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -6.15625,
+      "step": 505
+    },
+    {
+      "epoch": 0.6208588957055214,
+      "grad_norm": 32.33588195685234,
+      "learning_rate": 6.409917355371901e-07,
+      "logits/chosen": -0.08349609375,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2588,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.90625,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.875,
+      "step": 506
+    },
+    {
+      "epoch": 0.6220858895705521,
+      "grad_norm": 29.030680879470466,
+      "learning_rate": 6.406611570247934e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -308.0,
+      "loss": 0.1953,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.90234375,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -6.15625,
+      "step": 507
+    },
+    {
+      "epoch": 0.6233128834355828,
+      "grad_norm": 31.4626526695897,
+      "learning_rate": 6.403305785123967e-07,
+      "logits/chosen": -0.0201416015625,
+      "logits/rejected": -0.2021484375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -310.0,
+      "loss": 0.24,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.671875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.59375,
+      "step": 508
+    },
+    {
+      "epoch": 0.6245398773006134,
+      "grad_norm": 42.61520515559976,
+      "learning_rate": 6.4e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -466.0,
+      "logps/rejected": -384.0,
+      "loss": 0.3973,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8984375,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -6.0625,
+      "step": 509
+    },
+    {
+      "epoch": 0.6257668711656442,
+      "grad_norm": 30.165060750168916,
+      "learning_rate": 6.396694214876032e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2226,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.306640625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.625,
+      "step": 510
+    },
+    {
+      "epoch": 0.6269938650306749,
+      "grad_norm": 40.15670022385489,
+      "learning_rate": 6.393388429752065e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3348,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.81640625,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -6.1875,
+      "step": 511
+    },
+    {
+      "epoch": 0.6282208588957056,
+      "grad_norm": 44.290408519253646,
+      "learning_rate": 6.390082644628099e-07,
+      "logits/chosen": -0.09130859375,
+      "logits/rejected": -0.2431640625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -384.0,
+      "loss": 0.3621,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.35546875,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -5.0,
+      "step": 512
+    },
+    {
+      "epoch": 0.6294478527607362,
+      "grad_norm": 37.44250565144745,
+      "learning_rate": 6.386776859504132e-07,
+      "logits/chosen": -0.09521484375,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -312.0,
+      "loss": 0.337,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -5.375,
+      "step": 513
+    },
+    {
+      "epoch": 0.6306748466257669,
+      "grad_norm": 29.97245642529846,
+      "learning_rate": 6.383471074380165e-07,
+      "logits/chosen": -0.0830078125,
+      "logits/rejected": -0.2412109375,
+      "logps/chosen": -330.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2436,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": -0.051513671875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -6.03125,
+      "step": 514
+    },
+    {
+      "epoch": 0.6319018404907976,
+      "grad_norm": 36.989875573277935,
+      "learning_rate": 6.380165289256198e-07,
+      "logits/chosen": -0.09912109375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2882,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": -0.041748046875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.84375,
+      "step": 515
+    },
+    {
+      "epoch": 0.6331288343558282,
+      "grad_norm": 31.306859511596812,
+      "learning_rate": 6.376859504132231e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2432,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.1142578125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.84375,
+      "step": 516
+    },
+    {
+      "epoch": 0.6343558282208589,
+      "grad_norm": 43.67170294973452,
+      "learning_rate": 6.373553719008264e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3884,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.53125,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -5.03125,
+      "step": 517
+    },
+    {
+      "epoch": 0.6355828220858896,
+      "grad_norm": 29.354877414703584,
+      "learning_rate": 6.370247933884298e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -342.0,
+      "loss": 0.1963,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.34375,
+      "step": 518
+    },
+    {
+      "epoch": 0.6368098159509202,
+      "grad_norm": 40.61854272265548,
+      "learning_rate": 6.36694214876033e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -334.0,
+      "loss": 0.299,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.75390625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.5,
+      "step": 519
+    },
+    {
+      "epoch": 0.6380368098159509,
+      "grad_norm": 38.052325639181014,
+      "learning_rate": 6.363636363636363e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3436,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.6875,
+      "step": 520
+    },
+    {
+      "epoch": 0.6392638036809816,
+      "grad_norm": 32.37002227141111,
+      "learning_rate": 6.360330578512396e-07,
+      "logits/chosen": -0.03271484375,
+      "logits/rejected": -0.171875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2525,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.5,
+      "step": 521
+    },
+    {
+      "epoch": 0.6404907975460122,
+      "grad_norm": 40.41274539784558,
+      "learning_rate": 6.357024793388429e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -420.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3152,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.96875,
+      "step": 522
+    },
+    {
+      "epoch": 0.6417177914110429,
+      "grad_norm": 42.30362208986331,
+      "learning_rate": 6.353719008264462e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3605,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.6171875,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -5.03125,
+      "step": 523
+    },
+    {
+      "epoch": 0.6429447852760736,
+      "grad_norm": 46.893532525519284,
+      "learning_rate": 6.350413223140496e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3361,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.0625,
+      "step": 524
+    },
+    {
+      "epoch": 0.6441717791411042,
+      "grad_norm": 49.422180518954264,
+      "learning_rate": 6.347107438016529e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.185546875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3927,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.498046875,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -4.65625,
+      "step": 525
+    },
+    {
+      "epoch": 0.645398773006135,
+      "grad_norm": 38.5530601675623,
+      "learning_rate": 6.343801652892562e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2504,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.5625,
+      "step": 526
+    },
+    {
+      "epoch": 0.6466257668711657,
+      "grad_norm": 35.94650372227145,
+      "learning_rate": 6.340495867768595e-07,
+      "logits/chosen": -0.0859375,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2888,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.9609375,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.1875,
+      "step": 527
+    },
+    {
+      "epoch": 0.6478527607361964,
+      "grad_norm": 33.57585022481078,
+      "learning_rate": 6.337190082644627e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2986,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.71875,
+      "step": 528
+    },
+    {
+      "epoch": 0.649079754601227,
+      "grad_norm": 35.863933883867006,
+      "learning_rate": 6.33388429752066e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3657,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.1875,
+      "step": 529
+    },
+    {
+      "epoch": 0.6503067484662577,
+      "grad_norm": 41.86945425639466,
+      "learning_rate": 6.330578512396694e-07,
+      "logits/chosen": -0.12353515625,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3197,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.375,
+      "step": 530
+    },
+    {
+      "epoch": 0.6515337423312884,
+      "grad_norm": 33.3934976525138,
+      "learning_rate": 6.327272727272727e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2751,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -5.1875,
+      "step": 531
+    },
+    {
+      "epoch": 0.652760736196319,
+      "grad_norm": 35.72821914002383,
+      "learning_rate": 6.32396694214876e-07,
+      "logits/chosen": -0.07958984375,
+      "logits/rejected": -0.2197265625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2782,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.875,
+      "step": 532
+    },
+    {
+      "epoch": 0.6539877300613497,
+      "grad_norm": 36.05732236551947,
+      "learning_rate": 6.320661157024793e-07,
+      "logits/chosen": -0.2421875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2611,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.8671875,
+      "rewards/margins": 7.3125,
+      "rewards/rejected": -5.4375,
+      "step": 533
+    },
+    {
+      "epoch": 0.6552147239263804,
+      "grad_norm": 43.004226294889776,
+      "learning_rate": 6.317355371900826e-07,
+      "logits/chosen": -0.10302734375,
+      "logits/rejected": -0.228515625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2813,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.84375,
+      "step": 534
+    },
+    {
+      "epoch": 0.656441717791411,
+      "grad_norm": 39.5682500491772,
+      "learning_rate": 6.31404958677686e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2961,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.380859375,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.5625,
+      "step": 535
+    },
+    {
+      "epoch": 0.6576687116564417,
+      "grad_norm": 36.55745143954986,
+      "learning_rate": 6.310743801652893e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2275,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.4375,
+      "step": 536
+    },
+    {
+      "epoch": 0.6588957055214724,
+      "grad_norm": 45.10266990819029,
+      "learning_rate": 6.307438016528925e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3939,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.265625,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -4.84375,
+      "step": 537
+    },
+    {
+      "epoch": 0.660122699386503,
+      "grad_norm": 44.09458580737532,
+      "learning_rate": 6.304132231404958e-07,
+      "logits/chosen": -0.1123046875,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3783,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.84375,
+      "step": 538
+    },
+    {
+      "epoch": 0.6613496932515337,
+      "grad_norm": 36.16672414551654,
+      "learning_rate": 6.300826446280991e-07,
+      "logits/chosen": -0.11181640625,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -340.0,
+      "loss": 0.309,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.40625,
+      "step": 539
+    },
+    {
+      "epoch": 0.6625766871165644,
+      "grad_norm": 39.077710779231424,
+      "learning_rate": 6.297520661157024e-07,
+      "logits/chosen": -0.0908203125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3217,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 6.65625,
+      "rewards/rejected": -5.15625,
+      "step": 540
+    },
+    {
+      "epoch": 0.6638036809815951,
+      "grad_norm": 35.971169171679556,
+      "learning_rate": 6.294214876033058e-07,
+      "logits/chosen": -0.10986328125,
+      "logits/rejected": -0.2080078125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2834,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.75390625,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -4.8125,
+      "step": 541
+    },
+    {
+      "epoch": 0.6650306748466258,
+      "grad_norm": 34.168657597082216,
+      "learning_rate": 6.290909090909091e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2812,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.65625,
+      "step": 542
+    },
+    {
+      "epoch": 0.6662576687116565,
+      "grad_norm": 36.23729440291612,
+      "learning_rate": 6.287603305785124e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -444.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2622,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -4.78125,
+      "step": 543
+    },
+    {
+      "epoch": 0.6674846625766871,
+      "grad_norm": 45.238447000820585,
+      "learning_rate": 6.284297520661157e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3695,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.91015625,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.25,
+      "step": 544
+    },
+    {
+      "epoch": 0.6687116564417178,
+      "grad_norm": 43.061817496062964,
+      "learning_rate": 6.28099173553719e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -452.0,
+      "logps/rejected": -360.0,
+      "loss": 0.4047,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.5625,
+      "step": 545
+    },
+    {
+      "epoch": 0.6699386503067485,
+      "grad_norm": 40.8862609592416,
+      "learning_rate": 6.277685950413222e-07,
+      "logits/chosen": -0.10205078125,
+      "logits/rejected": -0.20703125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3402,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.09375,
+      "step": 546
+    },
+    {
+      "epoch": 0.6711656441717792,
+      "grad_norm": 32.99712535949366,
+      "learning_rate": 6.274380165289256e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2883,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.9375,
+      "step": 547
+    },
+    {
+      "epoch": 0.6723926380368098,
+      "grad_norm": 40.33054062017028,
+      "learning_rate": 6.271074380165289e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2583,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.71875,
+      "step": 548
+    },
+    {
+      "epoch": 0.6736196319018405,
+      "grad_norm": 37.72507019090557,
+      "learning_rate": 6.267768595041322e-07,
+      "logits/chosen": -0.11474609375,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.84375,
+      "step": 549
+    },
+    {
+      "epoch": 0.6748466257668712,
+      "grad_norm": 37.30173963897267,
+      "learning_rate": 6.264462809917355e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3323,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.59375,
+      "step": 550
+    },
+    {
+      "epoch": 0.6760736196319018,
+      "grad_norm": 33.242769441390706,
+      "learning_rate": 6.261157024793388e-07,
+      "logits/chosen": -0.0986328125,
+      "logits/rejected": -0.2265625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2814,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9296875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -4.59375,
+      "step": 551
+    },
+    {
+      "epoch": 0.6773006134969325,
+      "grad_norm": 35.53857403218536,
+      "learning_rate": 6.257851239669421e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3055,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.6875,
+      "step": 552
+    },
+    {
+      "epoch": 0.6785276073619632,
+      "grad_norm": 39.8186178694702,
+      "learning_rate": 6.254545454545455e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3556,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.8125,
+      "step": 553
+    },
+    {
+      "epoch": 0.6797546012269938,
+      "grad_norm": 34.396570445757696,
+      "learning_rate": 6.251239669421488e-07,
+      "logits/chosen": -0.08349609375,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -326.0,
+      "loss": 0.261,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.6875,
+      "step": 554
+    },
+    {
+      "epoch": 0.6809815950920245,
+      "grad_norm": 37.46750845748253,
+      "learning_rate": 6.24793388429752e-07,
+      "logits/chosen": 0.0196533203125,
+      "logits/rejected": -0.107421875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3763,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.65625,
+      "step": 555
+    },
+    {
+      "epoch": 0.6822085889570552,
+      "grad_norm": 28.75095679008507,
+      "learning_rate": 6.244628099173553e-07,
+      "logits/chosen": -0.10693359375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2294,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.0625,
+      "step": 556
+    },
+    {
+      "epoch": 0.6834355828220859,
+      "grad_norm": 32.12117772558378,
+      "learning_rate": 6.241322314049586e-07,
+      "logits/chosen": -0.11669921875,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2775,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -4.65625,
+      "step": 557
+    },
+    {
+      "epoch": 0.6846625766871166,
+      "grad_norm": 30.45940478111348,
+      "learning_rate": 6.238016528925619e-07,
+      "logits/chosen": -0.07861328125,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2442,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.8359375,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.875,
+      "step": 558
+    },
+    {
+      "epoch": 0.6858895705521473,
+      "grad_norm": 38.28759745643993,
+      "learning_rate": 6.234710743801653e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3234,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.25,
+      "step": 559
+    },
+    {
+      "epoch": 0.6871165644171779,
+      "grad_norm": 25.47174077723704,
+      "learning_rate": 6.231404958677686e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2016,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -6.03125,
+      "step": 560
+    },
+    {
+      "epoch": 0.6883435582822086,
+      "grad_norm": 28.61531386213562,
+      "learning_rate": 6.228099173553719e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -300.0,
+      "loss": 0.2021,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.6875,
+      "step": 561
+    },
+    {
+      "epoch": 0.6895705521472393,
+      "grad_norm": 30.528229373220828,
+      "learning_rate": 6.224793388429752e-07,
+      "logits/chosen": -0.032470703125,
+      "logits/rejected": -0.2333984375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2462,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.5,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.5625,
+      "step": 562
+    },
+    {
+      "epoch": 0.69079754601227,
+      "grad_norm": 34.031991776353316,
+      "learning_rate": 6.221487603305785e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2904,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 7.21875,
+      "rewards/rejected": -5.65625,
+      "step": 563
+    },
+    {
+      "epoch": 0.6920245398773006,
+      "grad_norm": 24.836619890425215,
+      "learning_rate": 6.218181818181817e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -334.0,
+      "loss": 0.224,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.34375,
+      "step": 564
+    },
+    {
+      "epoch": 0.6932515337423313,
+      "grad_norm": 35.09286389818353,
+      "learning_rate": 6.214876033057851e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3639,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.6015625,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.84375,
+      "step": 565
+    },
+    {
+      "epoch": 0.694478527607362,
+      "grad_norm": 33.71563020260373,
+      "learning_rate": 6.211570247933884e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2462,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.1875,
+      "step": 566
+    },
+    {
+      "epoch": 0.6957055214723926,
+      "grad_norm": 25.84647564741135,
+      "learning_rate": 6.208264462809917e-07,
+      "logits/chosen": -0.053955078125,
+      "logits/rejected": -0.1875,
+      "logps/chosen": -316.0,
+      "logps/rejected": -274.0,
+      "loss": 0.2061,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.359375,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -5.09375,
+      "step": 567
+    },
+    {
+      "epoch": 0.6969325153374233,
+      "grad_norm": 37.567177147049684,
+      "learning_rate": 6.20495867768595e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3064,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.21875,
+      "step": 568
+    },
+    {
+      "epoch": 0.698159509202454,
+      "grad_norm": 32.35117698678835,
+      "learning_rate": 6.201652892561983e-07,
+      "logits/chosen": -0.10302734375,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2508,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.125,
+      "step": 569
+    },
+    {
+      "epoch": 0.6993865030674846,
+      "grad_norm": 25.32334987337483,
+      "learning_rate": 6.198347107438017e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -338.0,
+      "loss": 0.1649,
+      "rewards/accuracies": 0.9453125,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.34375,
+      "step": 570
+    },
+    {
+      "epoch": 0.7006134969325153,
+      "grad_norm": 31.066523168854104,
+      "learning_rate": 6.19504132231405e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -326.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2262,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.6015625,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -5.09375,
+      "step": 571
+    },
+    {
+      "epoch": 0.701840490797546,
+      "grad_norm": 32.881821449548816,
+      "learning_rate": 6.191735537190083e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2291,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -4.9375,
+      "step": 572
+    },
+    {
+      "epoch": 0.7030674846625767,
+      "grad_norm": 40.68529761937124,
+      "learning_rate": 6.188429752066115e-07,
+      "logits/chosen": -0.07177734375,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4034,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -5.3125,
+      "step": 573
+    },
+    {
+      "epoch": 0.7042944785276074,
+      "grad_norm": 38.820291986225,
+      "learning_rate": 6.185123966942148e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -322.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2792,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.25,
+      "step": 574
+    },
+    {
+      "epoch": 0.7055214723926381,
+      "grad_norm": 33.59717634578736,
+      "learning_rate": 6.181818181818181e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2284,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.59375,
+      "step": 575
+    },
+    {
+      "epoch": 0.7067484662576687,
+      "grad_norm": 28.04916589516562,
+      "learning_rate": 6.178512396694215e-07,
+      "logits/chosen": -0.10302734375,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2001,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -5.875,
+      "step": 576
+    },
+    {
+      "epoch": 0.7079754601226994,
+      "grad_norm": 36.15869255614372,
+      "learning_rate": 6.175206611570248e-07,
+      "logits/chosen": -0.03955078125,
+      "logits/rejected": -0.1767578125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2929,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.84375,
+      "step": 577
+    },
+    {
+      "epoch": 0.7092024539877301,
+      "grad_norm": 36.89994352340565,
+      "learning_rate": 6.171900826446281e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -334.0,
+      "loss": 0.322,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.75,
+      "step": 578
+    },
+    {
+      "epoch": 0.7104294478527607,
+      "grad_norm": 42.55492037495249,
+      "learning_rate": 6.168595041322314e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3864,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.96875,
+      "step": 579
+    },
+    {
+      "epoch": 0.7116564417177914,
+      "grad_norm": 35.02868013813214,
+      "learning_rate": 6.165289256198347e-07,
+      "logits/chosen": -0.07275390625,
+      "logits/rejected": -0.2451171875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -354.0,
+      "loss": 0.229,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.73046875,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -6.25,
+      "step": 580
+    },
+    {
+      "epoch": 0.7128834355828221,
+      "grad_norm": 29.748844282456343,
+      "learning_rate": 6.16198347107438e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2071,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.84375,
+      "step": 581
+    },
+    {
+      "epoch": 0.7141104294478527,
+      "grad_norm": 41.261123421442086,
+      "learning_rate": 6.158677685950413e-07,
+      "logits/chosen": 0.01953125,
+      "logits/rejected": -0.197265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3132,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.78125,
+      "step": 582
+    },
+    {
+      "epoch": 0.7153374233128834,
+      "grad_norm": 37.14443132697755,
+      "learning_rate": 6.155371900826446e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3036,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.875,
+      "step": 583
+    },
+    {
+      "epoch": 0.7165644171779141,
+      "grad_norm": 42.810461239623855,
+      "learning_rate": 6.152066115702479e-07,
+      "logits/chosen": -0.0693359375,
+      "logits/rejected": -0.14453125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -370.0,
+      "loss": 0.3929,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 0.5859375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.25,
+      "step": 584
+    },
+    {
+      "epoch": 0.7177914110429447,
+      "grad_norm": 44.860424166016166,
+      "learning_rate": 6.148760330578512e-07,
+      "logits/chosen": 0.0206298828125,
+      "logits/rejected": -0.099609375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -322.0,
+      "loss": 0.4253,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.0625,
+      "step": 585
+    },
+    {
+      "epoch": 0.7190184049079754,
+      "grad_norm": 36.65848945690881,
+      "learning_rate": 6.145454545454545e-07,
+      "logits/chosen": -0.1064453125,
+      "logits/rejected": -0.2158203125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3226,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.5625,
+      "step": 586
+    },
+    {
+      "epoch": 0.7202453987730061,
+      "grad_norm": 34.66555869797275,
+      "learning_rate": 6.142148760330578e-07,
+      "logits/chosen": -0.1396484375,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -332.0,
+      "logps/rejected": -308.0,
+      "loss": 0.319,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.5,
+      "step": 587
+    },
+    {
+      "epoch": 0.7214723926380369,
+      "grad_norm": 25.81313251350675,
+      "learning_rate": 6.138842975206612e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2328,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9765625,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -4.34375,
+      "step": 588
+    },
+    {
+      "epoch": 0.7226993865030675,
+      "grad_norm": 39.45475953492192,
+      "learning_rate": 6.135537190082645e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3428,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.7734375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.0625,
+      "step": 589
+    },
+    {
+      "epoch": 0.7239263803680982,
+      "grad_norm": 28.739570926494295,
+      "learning_rate": 6.132231404958677e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.2392578125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2629,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -4.375,
+      "step": 590
+    },
+    {
+      "epoch": 0.7251533742331289,
+      "grad_norm": 39.542558184509424,
+      "learning_rate": 6.12892561983471e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2455,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -4.9375,
+      "step": 591
+    },
+    {
+      "epoch": 0.7263803680981595,
+      "grad_norm": 31.700565699717185,
+      "learning_rate": 6.125619834710743e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2372,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.1875,
+      "step": 592
+    },
+    {
+      "epoch": 0.7276073619631902,
+      "grad_norm": 43.160388567583965,
+      "learning_rate": 6.122314049586776e-07,
+      "logits/chosen": -0.08251953125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -442.0,
+      "logps/rejected": -366.0,
+      "loss": 0.3556,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -4.5625,
+      "step": 593
+    },
+    {
+      "epoch": 0.7288343558282209,
+      "grad_norm": 33.17336982955441,
+      "learning_rate": 6.11900826446281e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2737,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.59375,
+      "step": 594
+    },
+    {
+      "epoch": 0.7300613496932515,
+      "grad_norm": 35.78921531146286,
+      "learning_rate": 6.115702479338843e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -444.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2188,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.40625,
+      "step": 595
+    },
+    {
+      "epoch": 0.7312883435582822,
+      "grad_norm": 39.132566907776244,
+      "learning_rate": 6.112396694214876e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3258,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.125,
+      "step": 596
+    },
+    {
+      "epoch": 0.7325153374233129,
+      "grad_norm": 36.16852227530472,
+      "learning_rate": 6.109090909090909e-07,
+      "logits/chosen": -0.03515625,
+      "logits/rejected": -0.2138671875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3225,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.9375,
+      "step": 597
+    },
+    {
+      "epoch": 0.7337423312883435,
+      "grad_norm": 29.88526191086389,
+      "learning_rate": 6.105785123966942e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -438.0,
+      "logps/rejected": -318.0,
+      "loss": 0.204,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 7.21875,
+      "rewards/rejected": -5.125,
+      "step": 598
+    },
+    {
+      "epoch": 0.7349693251533742,
+      "grad_norm": 32.885582252909025,
+      "learning_rate": 6.102479338842974e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -448.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2847,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.53125,
+      "step": 599
+    },
+    {
+      "epoch": 0.7361963190184049,
+      "grad_norm": 37.722786017646094,
+      "learning_rate": 6.099173553719008e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -318.0,
+      "loss": 0.368,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.53125,
+      "step": 600
+    },
+    {
+      "epoch": 0.7374233128834355,
+      "grad_norm": 31.64488245915107,
+      "learning_rate": 6.095867768595041e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2698,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.1826171875,
+      "rewards/margins": 6.125,
+      "rewards/rejected": -5.9375,
+      "step": 601
+    },
+    {
+      "epoch": 0.7386503067484662,
+      "grad_norm": 25.124092304435703,
+      "learning_rate": 6.092561983471074e-07,
+      "logits/chosen": -0.0076904296875,
+      "logits/rejected": -0.1708984375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -332.0,
+      "loss": 0.189,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 7.28125,
+      "rewards/rejected": -6.03125,
+      "step": 602
+    },
+    {
+      "epoch": 0.7398773006134969,
+      "grad_norm": 30.45546682074925,
+      "learning_rate": 6.089256198347107e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2864,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.625,
+      "step": 603
+    },
+    {
+      "epoch": 0.7411042944785277,
+      "grad_norm": 31.94436328392251,
+      "learning_rate": 6.08595041322314e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2464,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.90625,
+      "step": 604
+    },
+    {
+      "epoch": 0.7423312883435583,
+      "grad_norm": 33.65745564504924,
+      "learning_rate": 6.082644628099173e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2598,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.984375,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -6.15625,
+      "step": 605
+    },
+    {
+      "epoch": 0.743558282208589,
+      "grad_norm": 38.63161562009038,
+      "learning_rate": 6.079338842975207e-07,
+      "logits/chosen": -0.2421875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3454,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.3125,
+      "step": 606
+    },
+    {
+      "epoch": 0.7447852760736197,
+      "grad_norm": 33.230355625984764,
+      "learning_rate": 6.07603305785124e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2611,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.48828125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -6.40625,
+      "step": 607
+    },
+    {
+      "epoch": 0.7460122699386503,
+      "grad_norm": 33.50647832359134,
+      "learning_rate": 6.072727272727272e-07,
+      "logits/chosen": -0.0216064453125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2913,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.099609375,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -5.6875,
+      "step": 608
+    },
+    {
+      "epoch": 0.747239263803681,
+      "grad_norm": 35.459553964773974,
+      "learning_rate": 6.069421487603305e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2856,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.54296875,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -6.21875,
+      "step": 609
+    },
+    {
+      "epoch": 0.7484662576687117,
+      "grad_norm": 31.662560042647854,
+      "learning_rate": 6.066115702479338e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -444.0,
+      "logps/rejected": -376.0,
+      "loss": 0.2452,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 7.46875,
+      "rewards/rejected": -6.375,
+      "step": 610
+    },
+    {
+      "epoch": 0.7496932515337423,
+      "grad_norm": 49.54193875289406,
+      "learning_rate": 6.062809917355371e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -362.0,
+      "loss": 0.3976,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -0.1396484375,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -6.0,
+      "step": 611
+    },
+    {
+      "epoch": 0.750920245398773,
+      "grad_norm": 39.062720193674714,
+      "learning_rate": 6.059504132231405e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3777,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.546875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.40625,
+      "step": 612
+    },
+    {
+      "epoch": 0.7521472392638037,
+      "grad_norm": 35.9253243800041,
+      "learning_rate": 6.056198347107438e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.2216796875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3394,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.10107421875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.96875,
+      "step": 613
+    },
+    {
+      "epoch": 0.7533742331288343,
+      "grad_norm": 54.139929990418025,
+      "learning_rate": 6.052892561983471e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3142,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.78125,
+      "step": 614
+    },
+    {
+      "epoch": 0.754601226993865,
+      "grad_norm": 37.247633297004576,
+      "learning_rate": 6.049586776859504e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3089,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.84765625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -6.0,
+      "step": 615
+    },
+    {
+      "epoch": 0.7558282208588957,
+      "grad_norm": 31.187596328354086,
+      "learning_rate": 6.046280991735537e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2528,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.10302734375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -5.46875,
+      "step": 616
+    },
+    {
+      "epoch": 0.7570552147239263,
+      "grad_norm": 39.842219715736135,
+      "learning_rate": 6.04297520661157e-07,
+      "logits/chosen": -0.125,
+      "logits/rejected": -0.208984375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3542,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.48828125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.3125,
+      "step": 617
+    },
+    {
+      "epoch": 0.758282208588957,
+      "grad_norm": 43.96609425022703,
+      "learning_rate": 6.039669421487603e-07,
+      "logits/chosen": -0.052734375,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3723,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.51953125,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.9375,
+      "step": 618
+    },
+    {
+      "epoch": 0.7595092024539877,
+      "grad_norm": 32.587670722826026,
+      "learning_rate": 6.036363636363636e-07,
+      "logits/chosen": -0.0791015625,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2331,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.90625,
+      "step": 619
+    },
+    {
+      "epoch": 0.7607361963190185,
+      "grad_norm": 44.41642388280846,
+      "learning_rate": 6.033057851239669e-07,
+      "logits/chosen": -0.0830078125,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3088,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.5234375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.21875,
+      "step": 620
+    },
+    {
+      "epoch": 0.7619631901840491,
+      "grad_norm": 34.02108964551065,
+      "learning_rate": 6.029752066115702e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -478.0,
+      "logps/rejected": -368.0,
+      "loss": 0.2661,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 7.3125,
+      "rewards/rejected": -6.125,
+      "step": 621
+    },
+    {
+      "epoch": 0.7631901840490798,
+      "grad_norm": 31.832438088278238,
+      "learning_rate": 6.026446280991735e-07,
+      "logits/chosen": -0.0888671875,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -342.0,
+      "loss": 0.236,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.72265625,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -6.03125,
+      "step": 622
+    },
+    {
+      "epoch": 0.7644171779141105,
+      "grad_norm": 38.89511615880052,
+      "learning_rate": 6.023140495867769e-07,
+      "logits/chosen": -0.1181640625,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -452.0,
+      "logps/rejected": -376.0,
+      "loss": 0.3167,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.4453125,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.46875,
+      "step": 623
+    },
+    {
+      "epoch": 0.7656441717791411,
+      "grad_norm": 32.19915482428781,
+      "learning_rate": 6.019834710743802e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2307,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 7.78125,
+      "rewards/rejected": -6.53125,
+      "step": 624
+    },
+    {
+      "epoch": 0.7668711656441718,
+      "grad_norm": 32.198949298871604,
+      "learning_rate": 6.016528925619835e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2191,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.375,
+      "step": 625
+    },
+    {
+      "epoch": 0.7680981595092025,
+      "grad_norm": 40.391116665536885,
+      "learning_rate": 6.013223140495867e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3346,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.03125,
+      "step": 626
+    },
+    {
+      "epoch": 0.7693251533742331,
+      "grad_norm": 33.50213686948866,
+      "learning_rate": 6.0099173553719e-07,
+      "logits/chosen": -0.06396484375,
+      "logits/rejected": -0.154296875,
+      "logps/chosen": -310.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3297,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": -0.1552734375,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -5.21875,
+      "step": 627
+    },
+    {
+      "epoch": 0.7705521472392638,
+      "grad_norm": 37.25258333430096,
+      "learning_rate": 6.006611570247933e-07,
+      "logits/chosen": -0.0947265625,
+      "logits/rejected": -0.2392578125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2886,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.90625,
+      "step": 628
+    },
+    {
+      "epoch": 0.7717791411042945,
+      "grad_norm": 23.68892864954692,
+      "learning_rate": 6.003305785123967e-07,
+      "logits/chosen": -0.1220703125,
+      "logits/rejected": -0.2470703125,
+      "logps/chosen": -348.0,
+      "logps/rejected": -284.0,
+      "loss": 0.1789,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.375,
+      "step": 629
+    },
+    {
+      "epoch": 0.7730061349693251,
+      "grad_norm": 33.10208379218177,
+      "learning_rate": 6e-07,
+      "logits/chosen": -0.060546875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2663,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.0625,
+      "step": 630
+    },
+    {
+      "epoch": 0.7742331288343558,
+      "grad_norm": 36.528138511875646,
+      "learning_rate": 5.996694214876033e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3321,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.875,
+      "step": 631
+    },
+    {
+      "epoch": 0.7754601226993865,
+      "grad_norm": 36.37506860284647,
+      "learning_rate": 5.993388429752066e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3056,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.34375,
+      "step": 632
+    },
+    {
+      "epoch": 0.7766871165644171,
+      "grad_norm": 40.1484232566858,
+      "learning_rate": 5.990082644628099e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.21484375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3418,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.84375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.65625,
+      "step": 633
+    },
+    {
+      "epoch": 0.7779141104294478,
+      "grad_norm": 32.83617762178678,
+      "learning_rate": 5.986776859504132e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2202,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 2.734375,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -5.1875,
+      "step": 634
+    },
+    {
+      "epoch": 0.7791411042944786,
+      "grad_norm": 29.099912658258905,
+      "learning_rate": 5.983471074380165e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2119,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.84375,
+      "step": 635
+    },
+    {
+      "epoch": 0.7803680981595092,
+      "grad_norm": 39.54581126371982,
+      "learning_rate": 5.980165289256198e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -366.0,
+      "loss": 0.3099,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -4.90625,
+      "step": 636
+    },
+    {
+      "epoch": 0.7815950920245399,
+      "grad_norm": 45.109718146173684,
+      "learning_rate": 5.976859504132231e-07,
+      "logits/chosen": -0.03125,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3522,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.0,
+      "step": 637
+    },
+    {
+      "epoch": 0.7828220858895706,
+      "grad_norm": 29.717665569657886,
+      "learning_rate": 5.973553719008264e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2213,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.9375,
+      "step": 638
+    },
+    {
+      "epoch": 0.7840490797546013,
+      "grad_norm": 38.75212680036792,
+      "learning_rate": 5.970247933884297e-07,
+      "logits/chosen": -0.048828125,
+      "logits/rejected": -0.2265625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3316,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.6484375,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.34375,
+      "step": 639
+    },
+    {
+      "epoch": 0.7852760736196319,
+      "grad_norm": 39.02043679792906,
+      "learning_rate": 5.96694214876033e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -440.0,
+      "logps/rejected": -382.0,
+      "loss": 0.2894,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -4.71875,
+      "step": 640
+    },
+    {
+      "epoch": 0.7865030674846626,
+      "grad_norm": 44.494628054471185,
+      "learning_rate": 5.963636363636364e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -348.0,
+      "loss": 0.4553,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.03125,
+      "step": 641
+    },
+    {
+      "epoch": 0.7877300613496933,
+      "grad_norm": 27.200186618621586,
+      "learning_rate": 5.960330578512397e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2487,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.5,
+      "step": 642
+    },
+    {
+      "epoch": 0.7889570552147239,
+      "grad_norm": 36.983988342062,
+      "learning_rate": 5.95702479338843e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -372.0,
+      "loss": 0.2631,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -4.53125,
+      "step": 643
+    },
+    {
+      "epoch": 0.7901840490797546,
+      "grad_norm": 39.431613095622076,
+      "learning_rate": 5.953719008264462e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2856,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.09375,
+      "step": 644
+    },
+    {
+      "epoch": 0.7914110429447853,
+      "grad_norm": 33.34429063414653,
+      "learning_rate": 5.950413223140495e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -328.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2445,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -4.9375,
+      "step": 645
+    },
+    {
+      "epoch": 0.7926380368098159,
+      "grad_norm": 34.57872089054598,
+      "learning_rate": 5.947107438016528e-07,
+      "logits/chosen": -0.0859375,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2881,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.75,
+      "step": 646
+    },
+    {
+      "epoch": 0.7938650306748466,
+      "grad_norm": 34.439204602754096,
+      "learning_rate": 5.943801652892562e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2819,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -5.125,
+      "step": 647
+    },
+    {
+      "epoch": 0.7950920245398773,
+      "grad_norm": 41.50373217977521,
+      "learning_rate": 5.940495867768595e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -324.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3682,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.859375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.84375,
+      "step": 648
+    },
+    {
+      "epoch": 0.7963190184049079,
+      "grad_norm": 43.045318488956525,
+      "learning_rate": 5.937190082644628e-07,
+      "logits/chosen": -0.0654296875,
+      "logits/rejected": -0.1650390625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.491,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.21875,
+      "step": 649
+    },
+    {
+      "epoch": 0.7975460122699386,
+      "grad_norm": 41.49765653672477,
+      "learning_rate": 5.933884297520661e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3786,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.09375,
+      "step": 650
+    },
+    {
+      "epoch": 0.7987730061349694,
+      "grad_norm": 38.92619113627064,
+      "learning_rate": 5.930578512396694e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2788,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -5.5625,
+      "step": 651
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 40.314158512080596,
+      "learning_rate": 5.927272727272728e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3015,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -5.09375,
+      "step": 652
+    },
+    {
+      "epoch": 0.8012269938650307,
+      "grad_norm": 41.91917001542575,
+      "learning_rate": 5.92396694214876e-07,
+      "logits/chosen": -0.103515625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2786,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.625,
+      "step": 653
+    },
+    {
+      "epoch": 0.8024539877300614,
+      "grad_norm": 32.50826629622854,
+      "learning_rate": 5.920661157024793e-07,
+      "logits/chosen": -0.083984375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2325,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.34375,
+      "step": 654
+    },
+    {
+      "epoch": 0.803680981595092,
+      "grad_norm": 36.66838919465664,
+      "learning_rate": 5.917355371900826e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2743,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.15625,
+      "step": 655
+    },
+    {
+      "epoch": 0.8049079754601227,
+      "grad_norm": 40.353892524687666,
+      "learning_rate": 5.914049586776859e-07,
+      "logits/chosen": -0.080078125,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3542,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -5.25,
+      "step": 656
+    },
+    {
+      "epoch": 0.8061349693251534,
+      "grad_norm": 32.71653938190321,
+      "learning_rate": 5.910743801652892e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3238,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.1875,
+      "step": 657
+    },
+    {
+      "epoch": 0.807361963190184,
+      "grad_norm": 35.63490892300594,
+      "learning_rate": 5.907438016528926e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2672,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.375,
+      "step": 658
+    },
+    {
+      "epoch": 0.8085889570552147,
+      "grad_norm": 36.34676950346043,
+      "learning_rate": 5.904132231404959e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3323,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.2177734375,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -5.6875,
+      "step": 659
+    },
+    {
+      "epoch": 0.8098159509202454,
+      "grad_norm": 31.25262331885713,
+      "learning_rate": 5.900826446280992e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2661,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.4296875,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.46875,
+      "step": 660
+    },
+    {
+      "epoch": 0.811042944785276,
+      "grad_norm": 44.04724360514248,
+      "learning_rate": 5.897520661157025e-07,
+      "logits/chosen": -0.11181640625,
+      "logits/rejected": -0.2236328125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3716,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.3125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.46875,
+      "step": 661
+    },
+    {
+      "epoch": 0.8122699386503067,
+      "grad_norm": 28.306859195207647,
+      "learning_rate": 5.894214876033057e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2117,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -5.71875,
+      "step": 662
+    },
+    {
+      "epoch": 0.8134969325153374,
+      "grad_norm": 28.755676702838585,
+      "learning_rate": 5.89090909090909e-07,
+      "logits/chosen": -0.06689453125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2615,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.09375,
+      "step": 663
+    },
+    {
+      "epoch": 0.8147239263803681,
+      "grad_norm": 30.20251070806522,
+      "learning_rate": 5.887603305785124e-07,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -358.0,
+      "loss": 0.1821,
+      "rewards/accuracies": 0.9140625,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 7.84375,
+      "rewards/rejected": -6.625,
+      "step": 664
+    },
+    {
+      "epoch": 0.8159509202453987,
+      "grad_norm": 43.59493072844384,
+      "learning_rate": 5.884297520661157e-07,
+      "logits/chosen": -0.05029296875,
+      "logits/rejected": -0.2197265625,
+      "logps/chosen": -334.0,
+      "logps/rejected": -292.0,
+      "loss": 0.2979,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.5234375,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.78125,
+      "step": 665
+    },
+    {
+      "epoch": 0.8171779141104294,
+      "grad_norm": 42.64325196007558,
+      "learning_rate": 5.88099173553719e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -348.0,
+      "loss": 0.368,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.21875,
+      "step": 666
+    },
+    {
+      "epoch": 0.8184049079754602,
+      "grad_norm": 34.27649786728366,
+      "learning_rate": 5.877685950413223e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2601,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.75,
+      "step": 667
+    },
+    {
+      "epoch": 0.8196319018404908,
+      "grad_norm": 33.27456675278737,
+      "learning_rate": 5.874380165289256e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2841,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.40625,
+      "step": 668
+    },
+    {
+      "epoch": 0.8208588957055215,
+      "grad_norm": 41.4043511264187,
+      "learning_rate": 5.871074380165289e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3017,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.375,
+      "step": 669
+    },
+    {
+      "epoch": 0.8220858895705522,
+      "grad_norm": 31.799223551777082,
+      "learning_rate": 5.867768595041323e-07,
+      "logits/chosen": -0.0191650390625,
+      "logits/rejected": -0.1650390625,
+      "logps/chosen": -324.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2541,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.125,
+      "step": 670
+    },
+    {
+      "epoch": 0.8233128834355828,
+      "grad_norm": 31.913109572694793,
+      "learning_rate": 5.864462809917355e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2407,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.0625,
+      "step": 671
+    },
+    {
+      "epoch": 0.8245398773006135,
+      "grad_norm": 32.62700830171627,
+      "learning_rate": 5.861157024793388e-07,
+      "logits/chosen": -0.109375,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -444.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2681,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.078125,
+      "rewards/margins": 7.25,
+      "rewards/rejected": -5.15625,
+      "step": 672
+    },
+    {
+      "epoch": 0.8257668711656442,
+      "grad_norm": 36.66995682138012,
+      "learning_rate": 5.857851239669421e-07,
+      "logits/chosen": -0.0498046875,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2592,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.625,
+      "step": 673
+    },
+    {
+      "epoch": 0.8269938650306748,
+      "grad_norm": 34.606200421728126,
+      "learning_rate": 5.854545454545454e-07,
+      "logits/chosen": -0.01300048828125,
+      "logits/rejected": -0.177734375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2822,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.09375,
+      "step": 674
+    },
+    {
+      "epoch": 0.8282208588957055,
+      "grad_norm": 35.038586845163366,
+      "learning_rate": 5.851239669421487e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3168,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.28125,
+      "step": 675
+    },
+    {
+      "epoch": 0.8294478527607362,
+      "grad_norm": 35.465018732105385,
+      "learning_rate": 5.847933884297521e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2534,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.71875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.53125,
+      "step": 676
+    },
+    {
+      "epoch": 0.8306748466257668,
+      "grad_norm": 36.537944400776084,
+      "learning_rate": 5.844628099173554e-07,
+      "logits/chosen": -0.06982421875,
+      "logits/rejected": -0.2451171875,
+      "logps/chosen": -334.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2916,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.84375,
+      "step": 677
+    },
+    {
+      "epoch": 0.8319018404907975,
+      "grad_norm": 33.168213995988694,
+      "learning_rate": 5.841322314049587e-07,
+      "logits/chosen": -0.1005859375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2768,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.25,
+      "step": 678
+    },
+    {
+      "epoch": 0.8331288343558282,
+      "grad_norm": 37.74716831610831,
+      "learning_rate": 5.838016528925619e-07,
+      "logits/chosen": -0.01220703125,
+      "logits/rejected": -0.1259765625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3364,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.8125,
+      "step": 679
+    },
+    {
+      "epoch": 0.8343558282208589,
+      "grad_norm": 34.79631984317766,
+      "learning_rate": 5.834710743801652e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -454.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2698,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -5.65625,
+      "step": 680
+    },
+    {
+      "epoch": 0.8355828220858895,
+      "grad_norm": 27.30848566312445,
+      "learning_rate": 5.831404958677685e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -324.0,
+      "loss": 0.1854,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -6.25,
+      "step": 681
+    },
+    {
+      "epoch": 0.8368098159509203,
+      "grad_norm": 36.85289178579729,
+      "learning_rate": 5.828099173553719e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3333,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.81640625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.65625,
+      "step": 682
+    },
+    {
+      "epoch": 0.838036809815951,
+      "grad_norm": 33.68839599065418,
+      "learning_rate": 5.824793388429752e-07,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2767,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.203125,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.8125,
+      "step": 683
+    },
+    {
+      "epoch": 0.8392638036809816,
+      "grad_norm": 35.217179760941754,
+      "learning_rate": 5.821487603305785e-07,
+      "logits/chosen": 0.000904083251953125,
+      "logits/rejected": -0.1181640625,
+      "logps/chosen": -314.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3304,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": -0.0341796875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -5.4375,
+      "step": 684
+    },
+    {
+      "epoch": 0.8404907975460123,
+      "grad_norm": 34.3907812005455,
+      "learning_rate": 5.818181818181818e-07,
+      "logits/chosen": -0.11865234375,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2858,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.9375,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -6.59375,
+      "step": 685
+    },
+    {
+      "epoch": 0.841717791411043,
+      "grad_norm": 29.480838482306872,
+      "learning_rate": 5.814876033057851e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2598,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.8125,
+      "step": 686
+    },
+    {
+      "epoch": 0.8429447852760736,
+      "grad_norm": 34.28684664510481,
+      "learning_rate": 5.811570247933884e-07,
+      "logits/chosen": -0.11767578125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2709,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.65625,
+      "step": 687
+    },
+    {
+      "epoch": 0.8441717791411043,
+      "grad_norm": 35.903679070126266,
+      "learning_rate": 5.808264462809917e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2748,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.07958984375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -6.125,
+      "step": 688
+    },
+    {
+      "epoch": 0.845398773006135,
+      "grad_norm": 37.11450186775036,
+      "learning_rate": 5.80495867768595e-07,
+      "logits/chosen": -0.12353515625,
+      "logits/rejected": -0.2197265625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3107,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.2109375,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.375,
+      "step": 689
+    },
+    {
+      "epoch": 0.8466257668711656,
+      "grad_norm": 49.86231404884124,
+      "learning_rate": 5.801652892561983e-07,
+      "logits/chosen": 0.004058837890625,
+      "logits/rejected": -0.1943359375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -350.0,
+      "loss": 0.4457,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.3125,
+      "step": 690
+    },
+    {
+      "epoch": 0.8478527607361963,
+      "grad_norm": 30.320356705626047,
+      "learning_rate": 5.798347107438016e-07,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2435,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.71875,
+      "step": 691
+    },
+    {
+      "epoch": 0.849079754601227,
+      "grad_norm": 43.95989415122109,
+      "learning_rate": 5.795041322314049e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -336.0,
+      "loss": 0.282,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -5.34375,
+      "step": 692
+    },
+    {
+      "epoch": 0.8503067484662576,
+      "grad_norm": 33.0652784939274,
+      "learning_rate": 5.791735537190082e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -442.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2934,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.203125,
+      "rewards/margins": 7.8125,
+      "rewards/rejected": -5.625,
+      "step": 693
+    },
+    {
+      "epoch": 0.8515337423312883,
+      "grad_norm": 46.044690743816275,
+      "learning_rate": 5.788429752066116e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -330.0,
+      "loss": 0.4005,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.5625,
+      "step": 694
+    },
+    {
+      "epoch": 0.852760736196319,
+      "grad_norm": 47.292538608776496,
+      "learning_rate": 5.785123966942149e-07,
+      "logits/chosen": -0.11376953125,
+      "logits/rejected": -0.208984375,
+      "logps/chosen": -328.0,
+      "logps/rejected": -302.0,
+      "loss": 0.4681,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 4.625,
+      "rewards/rejected": -3.90625,
+      "step": 695
+    },
+    {
+      "epoch": 0.8539877300613496,
+      "grad_norm": 38.43478974924762,
+      "learning_rate": 5.781818181818182e-07,
+      "logits/chosen": -0.07568359375,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3104,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.6875,
+      "step": 696
+    },
+    {
+      "epoch": 0.8552147239263803,
+      "grad_norm": 58.82691398412282,
+      "learning_rate": 5.778512396694214e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3128,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.140625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -4.78125,
+      "step": 697
+    },
+    {
+      "epoch": 0.8564417177914111,
+      "grad_norm": 37.608292873182215,
+      "learning_rate": 5.775206611570247e-07,
+      "logits/chosen": -0.080078125,
+      "logits/rejected": -0.216796875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2946,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -4.84375,
+      "step": 698
+    },
+    {
+      "epoch": 0.8576687116564418,
+      "grad_norm": 41.848029972370156,
+      "learning_rate": 5.77190082644628e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.2421875,
+      "logps/chosen": -334.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2843,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.91015625,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.71875,
+      "step": 699
+    },
+    {
+      "epoch": 0.8588957055214724,
+      "grad_norm": 39.920137771888676,
+      "learning_rate": 5.768595041322314e-07,
+      "logits/chosen": -0.054931640625,
+      "logits/rejected": -0.1640625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3259,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.34375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.25,
+      "step": 700
+    },
+    {
+      "epoch": 0.8601226993865031,
+      "grad_norm": 41.52681124295287,
+      "learning_rate": 5.765289256198347e-07,
+      "logits/chosen": -0.025146484375,
+      "logits/rejected": -0.138671875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -322.0,
+      "loss": 0.385,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.78125,
+      "step": 701
+    },
+    {
+      "epoch": 0.8613496932515338,
+      "grad_norm": 28.252991079387233,
+      "learning_rate": 5.76198347107438e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -330.0,
+      "loss": 0.1786,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -5.21875,
+      "step": 702
+    },
+    {
+      "epoch": 0.8625766871165644,
+      "grad_norm": 31.56795802217109,
+      "learning_rate": 5.758677685950413e-07,
+      "logits/chosen": -0.046142578125,
+      "logits/rejected": -0.1884765625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2909,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.3125,
+      "step": 703
+    },
+    {
+      "epoch": 0.8638036809815951,
+      "grad_norm": 34.11635728276886,
+      "learning_rate": 5.755371900826446e-07,
+      "logits/chosen": -0.1201171875,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -320.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3105,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 6.1875,
+      "rewards/rejected": -4.9375,
+      "step": 704
+    },
+    {
+      "epoch": 0.8650306748466258,
+      "grad_norm": 38.95077133588588,
+      "learning_rate": 5.75206611570248e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2767,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.265625,
+      "rewards/margins": 7.15625,
+      "rewards/rejected": -4.875,
+      "step": 705
+    },
+    {
+      "epoch": 0.8662576687116564,
+      "grad_norm": 35.267737415428456,
+      "learning_rate": 5.748760330578512e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3068,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.53125,
+      "step": 706
+    },
+    {
+      "epoch": 0.8674846625766871,
+      "grad_norm": 34.32550780952695,
+      "learning_rate": 5.745454545454545e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2683,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.25,
+      "step": 707
+    },
+    {
+      "epoch": 0.8687116564417178,
+      "grad_norm": 31.650191661242673,
+      "learning_rate": 5.742148760330578e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2379,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -4.875,
+      "step": 708
+    },
+    {
+      "epoch": 0.8699386503067484,
+      "grad_norm": 25.32301991794396,
+      "learning_rate": 5.738842975206611e-07,
+      "logits/chosen": -0.06640625,
+      "logits/rejected": -0.23046875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -340.0,
+      "loss": 0.185,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.9765625,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -5.53125,
+      "step": 709
+    },
+    {
+      "epoch": 0.8711656441717791,
+      "grad_norm": 33.22472192244502,
+      "learning_rate": 5.735537190082644e-07,
+      "logits/chosen": -0.017333984375,
+      "logits/rejected": -0.15234375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -312.0,
+      "loss": 0.299,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.625,
+      "step": 710
+    },
+    {
+      "epoch": 0.8723926380368098,
+      "grad_norm": 29.02267964751041,
+      "learning_rate": 5.732231404958678e-07,
+      "logits/chosen": -0.042724609375,
+      "logits/rejected": -0.236328125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2551,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.78125,
+      "step": 711
+    },
+    {
+      "epoch": 0.8736196319018404,
+      "grad_norm": 40.668984053435686,
+      "learning_rate": 5.728925619834711e-07,
+      "logits/chosen": -0.03857421875,
+      "logits/rejected": -0.1923828125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3587,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.375,
+      "step": 712
+    },
+    {
+      "epoch": 0.8748466257668711,
+      "grad_norm": 33.08084767770349,
+      "learning_rate": 5.725619834710744e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2665,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.15625,
+      "step": 713
+    },
+    {
+      "epoch": 0.8760736196319019,
+      "grad_norm": 36.45936464644564,
+      "learning_rate": 5.722314049586777e-07,
+      "logits/chosen": -0.0159912109375,
+      "logits/rejected": -0.2333984375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2938,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.75,
+      "step": 714
+    },
+    {
+      "epoch": 0.8773006134969326,
+      "grad_norm": 30.913206903477224,
+      "learning_rate": 5.719008264462809e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2684,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.53125,
+      "step": 715
+    },
+    {
+      "epoch": 0.8785276073619632,
+      "grad_norm": 41.616080953336535,
+      "learning_rate": 5.715702479338842e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3142,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.625,
+      "step": 716
+    },
+    {
+      "epoch": 0.8797546012269939,
+      "grad_norm": 34.18654539125182,
+      "learning_rate": 5.712396694214876e-07,
+      "logits/chosen": -0.09423828125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -276.0,
+      "loss": 0.2911,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.9296875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.3125,
+      "step": 717
+    },
+    {
+      "epoch": 0.8809815950920246,
+      "grad_norm": 28.09115304120503,
+      "learning_rate": 5.709090909090909e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2254,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.859375,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.0625,
+      "step": 718
+    },
+    {
+      "epoch": 0.8822085889570552,
+      "grad_norm": 41.08426714668054,
+      "learning_rate": 5.705785123966942e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3011,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.25,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -4.9375,
+      "step": 719
+    },
+    {
+      "epoch": 0.8834355828220859,
+      "grad_norm": 28.518925312083375,
+      "learning_rate": 5.702479338842975e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2223,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.265625,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -4.78125,
+      "step": 720
+    },
+    {
+      "epoch": 0.8846625766871166,
+      "grad_norm": 26.149570167864507,
+      "learning_rate": 5.699173553719008e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2073,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.890625,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -6.09375,
+      "step": 721
+    },
+    {
+      "epoch": 0.8858895705521472,
+      "grad_norm": 30.929567326750046,
+      "learning_rate": 5.695867768595041e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2229,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.859375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.125,
+      "step": 722
+    },
+    {
+      "epoch": 0.8871165644171779,
+      "grad_norm": 34.37293415398814,
+      "learning_rate": 5.692561983471075e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2808,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -5.4375,
+      "step": 723
+    },
+    {
+      "epoch": 0.8883435582822086,
+      "grad_norm": 33.90677929130864,
+      "learning_rate": 5.689256198347107e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -308.0,
+      "loss": 0.254,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.7265625,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.21875,
+      "step": 724
+    },
+    {
+      "epoch": 0.8895705521472392,
+      "grad_norm": 34.97307578381243,
+      "learning_rate": 5.68595041322314e-07,
+      "logits/chosen": -0.055419921875,
+      "logits/rejected": -0.2314453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3044,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.15625,
+      "step": 725
+    },
+    {
+      "epoch": 0.8907975460122699,
+      "grad_norm": 28.627835648582934,
+      "learning_rate": 5.682644628099173e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -336.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2135,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.9765625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.5625,
+      "step": 726
+    },
+    {
+      "epoch": 0.8920245398773006,
+      "grad_norm": 34.81272514902735,
+      "learning_rate": 5.679338842975206e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -324.0,
+      "loss": 0.248,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.15625,
+      "step": 727
+    },
+    {
+      "epoch": 0.8932515337423312,
+      "grad_norm": 37.8936877345969,
+      "learning_rate": 5.676033057851239e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -398.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2946,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 7.65625,
+      "rewards/rejected": -5.625,
+      "step": 728
+    },
+    {
+      "epoch": 0.894478527607362,
+      "grad_norm": 29.90567314985844,
+      "learning_rate": 5.672727272727273e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2088,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 7.6875,
+      "rewards/rejected": -6.0625,
+      "step": 729
+    },
+    {
+      "epoch": 0.8957055214723927,
+      "grad_norm": 31.408249750098918,
+      "learning_rate": 5.669421487603306e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -398.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2204,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -5.75,
+      "step": 730
+    },
+    {
+      "epoch": 0.8969325153374234,
+      "grad_norm": 33.91624972638253,
+      "learning_rate": 5.666115702479339e-07,
+      "logits/chosen": -0.052734375,
+      "logits/rejected": -0.1845703125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2736,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.8125,
+      "step": 731
+    },
+    {
+      "epoch": 0.898159509202454,
+      "grad_norm": 40.7773464371931,
+      "learning_rate": 5.662809917355372e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3647,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.375,
+      "step": 732
+    },
+    {
+      "epoch": 0.8993865030674847,
+      "grad_norm": 31.001224216873553,
+      "learning_rate": 5.659504132231404e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2212,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.84375,
+      "step": 733
+    },
+    {
+      "epoch": 0.9006134969325154,
+      "grad_norm": 37.803749225628884,
+      "learning_rate": 5.656198347107437e-07,
+      "logits/chosen": -0.08837890625,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2954,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 7.5,
+      "rewards/rejected": -6.25,
+      "step": 734
+    },
+    {
+      "epoch": 0.901840490797546,
+      "grad_norm": 30.18047116909555,
+      "learning_rate": 5.652892561983471e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -372.0,
+      "loss": 0.1824,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -6.8125,
+      "step": 735
+    },
+    {
+      "epoch": 0.9030674846625767,
+      "grad_norm": 38.572832638835585,
+      "learning_rate": 5.649586776859504e-07,
+      "logits/chosen": -0.025634765625,
+      "logits/rejected": -0.1552734375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3608,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.0908203125,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -6.3125,
+      "step": 736
+    },
+    {
+      "epoch": 0.9042944785276074,
+      "grad_norm": 46.258369054797996,
+      "learning_rate": 5.646280991735537e-07,
+      "logits/chosen": -0.12451171875,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4383,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.1826171875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -6.25,
+      "step": 737
+    },
+    {
+      "epoch": 0.905521472392638,
+      "grad_norm": 32.39664225178629,
+      "learning_rate": 5.64297520661157e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -442.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2588,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 7.5625,
+      "rewards/rejected": -6.28125,
+      "step": 738
+    },
+    {
+      "epoch": 0.9067484662576687,
+      "grad_norm": 34.18126494383032,
+      "learning_rate": 5.639669421487603e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -344.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2836,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.2470703125,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -6.15625,
+      "step": 739
+    },
+    {
+      "epoch": 0.9079754601226994,
+      "grad_norm": 35.15083279583394,
+      "learning_rate": 5.636363636363636e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -342.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3072,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.2431640625,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -6.6875,
+      "step": 740
+    },
+    {
+      "epoch": 0.90920245398773,
+      "grad_norm": 33.535600643328806,
+      "learning_rate": 5.63305785123967e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2851,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -6.8125,
+      "step": 741
+    },
+    {
+      "epoch": 0.9104294478527607,
+      "grad_norm": 31.52809779595991,
+      "learning_rate": 5.629752066115702e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2244,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.90625,
+      "step": 742
+    },
+    {
+      "epoch": 0.9116564417177914,
+      "grad_norm": 35.37117075886455,
+      "learning_rate": 5.626446280991735e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -380.0,
+      "loss": 0.2769,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.8125,
+      "step": 743
+    },
+    {
+      "epoch": 0.912883435582822,
+      "grad_norm": 26.42035749866701,
+      "learning_rate": 5.623140495867768e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -330.0,
+      "loss": 0.1917,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.75,
+      "step": 744
+    },
+    {
+      "epoch": 0.9141104294478528,
+      "grad_norm": 35.853268232755404,
+      "learning_rate": 5.619834710743801e-07,
+      "logits/chosen": -0.08251953125,
+      "logits/rejected": -0.1865234375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3009,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.390625,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.875,
+      "step": 745
+    },
+    {
+      "epoch": 0.9153374233128835,
+      "grad_norm": 43.35686834196019,
+      "learning_rate": 5.616528925619834e-07,
+      "logits/chosen": -0.0849609375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -346.0,
+      "loss": 0.317,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -6.0,
+      "step": 746
+    },
+    {
+      "epoch": 0.9165644171779141,
+      "grad_norm": 34.17444251605582,
+      "learning_rate": 5.613223140495868e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2811,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.41796875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.78125,
+      "step": 747
+    },
+    {
+      "epoch": 0.9177914110429448,
+      "grad_norm": 40.3104413419788,
+      "learning_rate": 5.609917355371901e-07,
+      "logits/chosen": -0.05224609375,
+      "logits/rejected": -0.169921875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3882,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.0279541015625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -5.625,
+      "step": 748
+    },
+    {
+      "epoch": 0.9190184049079755,
+      "grad_norm": 40.57709610969512,
+      "learning_rate": 5.606611570247934e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3642,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.98828125,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -6.09375,
+      "step": 749
+    },
+    {
+      "epoch": 0.9202453987730062,
+      "grad_norm": 43.682412563973024,
+      "learning_rate": 5.603305785123967e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3704,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.63671875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.875,
+      "step": 750
+    },
+    {
+      "epoch": 0.9214723926380368,
+      "grad_norm": 42.01608666567515,
+      "learning_rate": 5.599999999999999e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3006,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 7.4375,
+      "rewards/rejected": -6.28125,
+      "step": 751
+    },
+    {
+      "epoch": 0.9226993865030675,
+      "grad_norm": 32.96496448782055,
+      "learning_rate": 5.596694214876033e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2435,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 7.75,
+      "rewards/rejected": -6.34375,
+      "step": 752
+    },
+    {
+      "epoch": 0.9239263803680982,
+      "grad_norm": 35.77444345136624,
+      "learning_rate": 5.593388429752066e-07,
+      "logits/chosen": -0.048095703125,
+      "logits/rejected": -0.1923828125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2851,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.796875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.5625,
+      "step": 753
+    },
+    {
+      "epoch": 0.9251533742331288,
+      "grad_norm": 39.56695959824003,
+      "learning_rate": 5.590082644628099e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3588,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.59375,
+      "step": 754
+    },
+    {
+      "epoch": 0.9263803680981595,
+      "grad_norm": 33.47777157003214,
+      "learning_rate": 5.586776859504132e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2421,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.6875,
+      "step": 755
+    },
+    {
+      "epoch": 0.9276073619631902,
+      "grad_norm": 30.73963254337421,
+      "learning_rate": 5.583471074380165e-07,
+      "logits/chosen": -0.0732421875,
+      "logits/rejected": -0.1669921875,
+      "logps/chosen": -332.0,
+      "logps/rejected": -308.0,
+      "loss": 0.248,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.322265625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.75,
+      "step": 756
+    },
+    {
+      "epoch": 0.9288343558282208,
+      "grad_norm": 25.863947966401952,
+      "learning_rate": 5.580165289256198e-07,
+      "logits/chosen": -0.0252685546875,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2117,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.375,
+      "step": 757
+    },
+    {
+      "epoch": 0.9300613496932515,
+      "grad_norm": 40.231833773135115,
+      "learning_rate": 5.576859504132232e-07,
+      "logits/chosen": -0.08837890625,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3497,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.46875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.59375,
+      "step": 758
+    },
+    {
+      "epoch": 0.9312883435582822,
+      "grad_norm": 34.14796678847348,
+      "learning_rate": 5.573553719008265e-07,
+      "logits/chosen": -0.0595703125,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2511,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.34375,
+      "step": 759
+    },
+    {
+      "epoch": 0.9325153374233128,
+      "grad_norm": 29.867916639384045,
+      "learning_rate": 5.570247933884297e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2724,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 7.8125,
+      "rewards/rejected": -5.71875,
+      "step": 760
+    },
+    {
+      "epoch": 0.9337423312883436,
+      "grad_norm": 38.414301746929866,
+      "learning_rate": 5.56694214876033e-07,
+      "logits/chosen": -0.0264892578125,
+      "logits/rejected": -0.1669921875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2874,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.375,
+      "step": 761
+    },
+    {
+      "epoch": 0.9349693251533743,
+      "grad_norm": 40.447076836962,
+      "learning_rate": 5.563636363636363e-07,
+      "logits/chosen": -0.01397705078125,
+      "logits/rejected": -0.19921875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2888,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.6875,
+      "step": 762
+    },
+    {
+      "epoch": 0.9361963190184049,
+      "grad_norm": 31.879754151772115,
+      "learning_rate": 5.560330578512396e-07,
+      "logits/chosen": -0.07373046875,
+      "logits/rejected": -0.171875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2711,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.4375,
+      "step": 763
+    },
+    {
+      "epoch": 0.9374233128834356,
+      "grad_norm": 35.145542304507835,
+      "learning_rate": 5.55702479338843e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -322.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2896,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.26171875,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -6.0,
+      "step": 764
+    },
+    {
+      "epoch": 0.9386503067484663,
+      "grad_norm": 40.67019743057879,
+      "learning_rate": 5.553719008264463e-07,
+      "logits/chosen": -0.0595703125,
+      "logits/rejected": -0.19140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3457,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.76171875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.96875,
+      "step": 765
+    },
+    {
+      "epoch": 0.939877300613497,
+      "grad_norm": 32.24931969694431,
+      "learning_rate": 5.550413223140496e-07,
+      "logits/chosen": -0.11767578125,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -326.0,
+      "loss": 0.274,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.5625,
+      "step": 766
+    },
+    {
+      "epoch": 0.9411042944785276,
+      "grad_norm": 34.46831261260643,
+      "learning_rate": 5.547107438016529e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3035,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.953125,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.9375,
+      "step": 767
+    },
+    {
+      "epoch": 0.9423312883435583,
+      "grad_norm": 36.15353177119284,
+      "learning_rate": 5.543801652892561e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -372.0,
+      "loss": 0.316,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -5.75,
+      "step": 768
+    },
+    {
+      "epoch": 0.943558282208589,
+      "grad_norm": 29.41748843520645,
+      "learning_rate": 5.540495867768594e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2432,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.6328125,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -6.375,
+      "step": 769
+    },
+    {
+      "epoch": 0.9447852760736196,
+      "grad_norm": 41.73128837710126,
+      "learning_rate": 5.537190082644628e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3004,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -5.46875,
+      "step": 770
+    },
+    {
+      "epoch": 0.9460122699386503,
+      "grad_norm": 42.981692808778455,
+      "learning_rate": 5.533884297520661e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.201171875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -346.0,
+      "loss": 0.429,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 0.0869140625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -5.5625,
+      "step": 771
+    },
+    {
+      "epoch": 0.947239263803681,
+      "grad_norm": 35.00554577923317,
+      "learning_rate": 5.530578512396694e-07,
+      "logits/chosen": -0.038330078125,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3114,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.6875,
+      "step": 772
+    },
+    {
+      "epoch": 0.9484662576687116,
+      "grad_norm": 36.88335000648106,
+      "learning_rate": 5.527272727272727e-07,
+      "logits/chosen": -0.0361328125,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3175,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.1513671875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.78125,
+      "step": 773
+    },
+    {
+      "epoch": 0.9496932515337423,
+      "grad_norm": 38.52772066032803,
+      "learning_rate": 5.52396694214876e-07,
+      "logits/chosen": -0.058837890625,
+      "logits/rejected": -0.2001953125,
+      "logps/chosen": -352.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3665,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": -0.330078125,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -5.0625,
+      "step": 774
+    },
+    {
+      "epoch": 0.950920245398773,
+      "grad_norm": 38.30272199053397,
+      "learning_rate": 5.520661157024793e-07,
+      "logits/chosen": 0.0306396484375,
+      "logits/rejected": -0.162109375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3245,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.03125,
+      "step": 775
+    },
+    {
+      "epoch": 0.9521472392638037,
+      "grad_norm": 33.16873912890908,
+      "learning_rate": 5.517355371900827e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3113,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -5.375,
+      "step": 776
+    },
+    {
+      "epoch": 0.9533742331288344,
+      "grad_norm": 35.432794054015964,
+      "learning_rate": 5.514049586776859e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3098,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.8125,
+      "step": 777
+    },
+    {
+      "epoch": 0.9546012269938651,
+      "grad_norm": 35.189286152943,
+      "learning_rate": 5.510743801652892e-07,
+      "logits/chosen": -0.1083984375,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -314.0,
+      "loss": 0.251,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.443359375,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -5.34375,
+      "step": 778
+    },
+    {
+      "epoch": 0.9558282208588957,
+      "grad_norm": 41.96194349470727,
+      "learning_rate": 5.507438016528925e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2827,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.96875,
+      "step": 779
+    },
+    {
+      "epoch": 0.9570552147239264,
+      "grad_norm": 35.73940962793884,
+      "learning_rate": 5.504132231404958e-07,
+      "logits/chosen": -0.059326171875,
+      "logits/rejected": -0.2353515625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3415,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.2138671875,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -5.4375,
+      "step": 780
+    },
+    {
+      "epoch": 0.9582822085889571,
+      "grad_norm": 26.50283260922046,
+      "learning_rate": 5.500826446280991e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -342.0,
+      "loss": 0.1906,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.4375,
+      "step": 781
+    },
+    {
+      "epoch": 0.9595092024539877,
+      "grad_norm": 29.07388144480791,
+      "learning_rate": 5.497520661157025e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2204,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.5625,
+      "step": 782
+    },
+    {
+      "epoch": 0.9607361963190184,
+      "grad_norm": 36.281519570454215,
+      "learning_rate": 5.494214876033058e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2544,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -6.09375,
+      "step": 783
+    },
+    {
+      "epoch": 0.9619631901840491,
+      "grad_norm": 41.7797714543972,
+      "learning_rate": 5.490909090909091e-07,
+      "logits/chosen": -0.11181640625,
+      "logits/rejected": -0.162109375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3585,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.1767578125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.625,
+      "step": 784
+    },
+    {
+      "epoch": 0.9631901840490797,
+      "grad_norm": 31.81554374340158,
+      "learning_rate": 5.487603305785124e-07,
+      "logits/chosen": -0.0751953125,
+      "logits/rejected": -0.24609375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2472,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.88671875,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -5.8125,
+      "step": 785
+    },
+    {
+      "epoch": 0.9644171779141104,
+      "grad_norm": 27.670018917456815,
+      "learning_rate": 5.484297520661156e-07,
+      "logits/chosen": -0.0732421875,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2222,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 0.4921875,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -6.25,
+      "step": 786
+    },
+    {
+      "epoch": 0.9656441717791411,
+      "grad_norm": 37.16124889650476,
+      "learning_rate": 5.480991735537189e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -6.4375,
+      "step": 787
+    },
+    {
+      "epoch": 0.9668711656441717,
+      "grad_norm": 24.391962735004444,
+      "learning_rate": 5.477685950413223e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -352.0,
+      "loss": 0.1706,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -6.25,
+      "step": 788
+    },
+    {
+      "epoch": 0.9680981595092024,
+      "grad_norm": 32.024862187416,
+      "learning_rate": 5.474380165289256e-07,
+      "logits/chosen": -0.07958984375,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2843,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.408203125,
+      "rewards/margins": 6.1875,
+      "rewards/rejected": -5.78125,
+      "step": 789
+    },
+    {
+      "epoch": 0.9693251533742331,
+      "grad_norm": 27.51754897072937,
+      "learning_rate": 5.471074380165289e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -316.0,
+      "loss": 0.1866,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -6.03125,
+      "step": 790
+    },
+    {
+      "epoch": 0.9705521472392638,
+      "grad_norm": 37.440979455018265,
+      "learning_rate": 5.467768595041322e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2983,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 7.59375,
+      "rewards/rejected": -6.375,
+      "step": 791
+    },
+    {
+      "epoch": 0.9717791411042945,
+      "grad_norm": 36.704827910830204,
+      "learning_rate": 5.464462809917355e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -430.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3431,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.03125,
+      "step": 792
+    },
+    {
+      "epoch": 0.9730061349693252,
+      "grad_norm": 37.60757704339766,
+      "learning_rate": 5.461157024793389e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3237,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.4375,
+      "step": 793
+    },
+    {
+      "epoch": 0.9742331288343559,
+      "grad_norm": 37.59792190871499,
+      "learning_rate": 5.457851239669422e-07,
+      "logits/chosen": -0.103515625,
+      "logits/rejected": -0.2431640625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3909,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.21875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -5.78125,
+      "step": 794
+    },
+    {
+      "epoch": 0.9754601226993865,
+      "grad_norm": 28.85691430451381,
+      "learning_rate": 5.454545454545454e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2394,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.90625,
+      "step": 795
+    },
+    {
+      "epoch": 0.9766871165644172,
+      "grad_norm": 41.11938718622486,
+      "learning_rate": 5.451239669421487e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2984,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": -0.0673828125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -6.375,
+      "step": 796
+    },
+    {
+      "epoch": 0.9779141104294479,
+      "grad_norm": 30.942308293626358,
+      "learning_rate": 5.44793388429752e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -308.0,
+      "loss": 0.254,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.625,
+      "step": 797
+    },
+    {
+      "epoch": 0.9791411042944785,
+      "grad_norm": 36.786308023977305,
+      "learning_rate": 5.444628099173553e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2823,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.76953125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -6.125,
+      "step": 798
+    },
+    {
+      "epoch": 0.9803680981595092,
+      "grad_norm": 40.89607573095922,
+      "learning_rate": 5.441322314049587e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3222,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.78125,
+      "step": 799
+    },
+    {
+      "epoch": 0.9815950920245399,
+      "grad_norm": 36.10519822526459,
+      "learning_rate": 5.43801652892562e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2658,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.61328125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -6.0,
+      "step": 800
+    },
+    {
+      "epoch": 0.9828220858895705,
+      "grad_norm": 42.48783397210697,
+      "learning_rate": 5.434710743801653e-07,
+      "logits/chosen": -0.17578125,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -366.0,
+      "loss": 0.3944,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.1875,
+      "step": 801
+    },
+    {
+      "epoch": 0.9840490797546012,
+      "grad_norm": 43.552763355211724,
+      "learning_rate": 5.431404958677686e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3899,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.453125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.84375,
+      "step": 802
+    },
+    {
+      "epoch": 0.9852760736196319,
+      "grad_norm": 32.09849729457811,
+      "learning_rate": 5.428099173553719e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2997,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.78125,
+      "step": 803
+    },
+    {
+      "epoch": 0.9865030674846625,
+      "grad_norm": 32.33615160143916,
+      "learning_rate": 5.424793388429751e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.240234375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3208,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.251953125,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -5.65625,
+      "step": 804
+    },
+    {
+      "epoch": 0.9877300613496932,
+      "grad_norm": 39.61512778302868,
+      "learning_rate": 5.421487603305785e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -380.0,
+      "loss": 0.3968,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.96875,
+      "step": 805
+    },
+    {
+      "epoch": 0.9889570552147239,
+      "grad_norm": 40.74080836640833,
+      "learning_rate": 5.418181818181818e-07,
+      "logits/chosen": -0.01202392578125,
+      "logits/rejected": -0.1474609375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -368.0,
+      "loss": 0.3811,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.34375,
+      "step": 806
+    },
+    {
+      "epoch": 0.9901840490797545,
+      "grad_norm": 36.652039792723684,
+      "learning_rate": 5.414876033057851e-07,
+      "logits/chosen": -0.1005859375,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2327,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 7.21875,
+      "rewards/rejected": -6.125,
+      "step": 807
+    },
+    {
+      "epoch": 0.9914110429447853,
+      "grad_norm": 33.57961890786831,
+      "learning_rate": 5.411570247933884e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3091,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.11572265625,
+      "rewards/margins": 6.65625,
+      "rewards/rejected": -6.53125,
+      "step": 808
+    },
+    {
+      "epoch": 0.992638036809816,
+      "grad_norm": 35.60132202210017,
+      "learning_rate": 5.408264462809917e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2618,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 7.875,
+      "rewards/rejected": -6.28125,
+      "step": 809
+    },
+    {
+      "epoch": 0.9938650306748467,
+      "grad_norm": 33.977380429457895,
+      "learning_rate": 5.40495867768595e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -368.0,
+      "loss": 0.227,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 7.28125,
+      "rewards/rejected": -6.0625,
+      "step": 810
+    },
+    {
+      "epoch": 0.9950920245398773,
+      "grad_norm": 31.635546850530186,
+      "learning_rate": 5.401652892561984e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2741,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.71875,
+      "step": 811
+    },
+    {
+      "epoch": 0.996319018404908,
+      "grad_norm": 40.80563506539147,
+      "learning_rate": 5.398347107438017e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2687,
+      "rewards/accuracies": 0.9140625,
+      "rewards/chosen": 0.4609375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -6.53125,
+      "step": 812
+    },
+    {
+      "epoch": 0.9975460122699387,
+      "grad_norm": 30.19996782105712,
+      "learning_rate": 5.395041322314049e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -338.0,
+      "logps/rejected": -324.0,
+      "loss": 0.1913,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.87109375,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -6.1875,
+      "step": 813
+    },
+    {
+      "epoch": 0.9987730061349693,
+      "grad_norm": 37.73036986625575,
+      "learning_rate": 5.391735537190082e-07,
+      "logits/chosen": -0.11767578125,
+      "logits/rejected": -0.201171875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3425,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": -0.1416015625,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -5.65625,
+      "step": 814
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 40.870567834511405,
+      "learning_rate": 5.388429752066115e-07,
+      "logits/chosen": -0.0458984375,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2125,
+      "rewards/accuracies": 0.8571429252624512,
+      "rewards/chosen": 0.59375,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -6.28125,
+      "step": 815
+    },
+    {
+      "epoch": 1.0012269938650307,
+      "grad_norm": 8.447732700524874,
+      "learning_rate": 5.385123966942148e-07,
+      "logits/chosen": -0.099609375,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -318.0,
+      "loss": 0.0478,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.75,
+      "step": 816
+    },
+    {
+      "epoch": 1.0024539877300613,
+      "grad_norm": 10.238235833774846,
+      "learning_rate": 5.381818181818182e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -320.0,
+      "loss": 0.0599,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.8828125,
+      "rewards/margins": 7.59375,
+      "rewards/rejected": -6.71875,
+      "step": 817
+    },
+    {
+      "epoch": 1.003680981595092,
+      "grad_norm": 8.509110400177521,
+      "learning_rate": 5.378512396694215e-07,
+      "logits/chosen": -0.248046875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0469,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -6.53125,
+      "step": 818
+    },
+    {
+      "epoch": 1.0049079754601227,
+      "grad_norm": 11.32230173464338,
+      "learning_rate": 5.375206611570248e-07,
+      "logits/chosen": -0.1083984375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0506,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -6.65625,
+      "step": 819
+    },
+    {
+      "epoch": 1.0061349693251533,
+      "grad_norm": 9.131424372223067,
+      "learning_rate": 5.371900826446281e-07,
+      "logits/chosen": -0.0751953125,
+      "logits/rejected": -0.216796875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0535,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.140625,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -5.78125,
+      "step": 820
+    },
+    {
+      "epoch": 1.007361963190184,
+      "grad_norm": 16.500866897905247,
+      "learning_rate": 5.368595041322314e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -344.0,
+      "logps/rejected": -326.0,
+      "loss": 0.083,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.0625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.4375,
+      "step": 821
+    },
+    {
+      "epoch": 1.0085889570552147,
+      "grad_norm": 10.055056560721704,
+      "learning_rate": 5.365289256198346e-07,
+      "logits/chosen": -0.1513671875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0514,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.9765625,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -6.03125,
+      "step": 822
+    },
+    {
+      "epoch": 1.0098159509202453,
+      "grad_norm": 8.750164950681329,
+      "learning_rate": 5.36198347107438e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0496,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.265625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.21875,
+      "step": 823
+    },
+    {
+      "epoch": 1.011042944785276,
+      "grad_norm": 8.95759785702542,
+      "learning_rate": 5.358677685950413e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0275,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.203125,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -5.9375,
+      "step": 824
+    },
+    {
+      "epoch": 1.0122699386503067,
+      "grad_norm": 8.774577043038853,
+      "learning_rate": 5.355371900826446e-07,
+      "logits/chosen": -0.021484375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0431,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -6.3125,
+      "step": 825
+    },
+    {
+      "epoch": 1.0134969325153373,
+      "grad_norm": 13.971484339500186,
+      "learning_rate": 5.352066115702479e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0773,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 7.375,
+      "rewards/rejected": -5.78125,
+      "step": 826
+    },
+    {
+      "epoch": 1.014723926380368,
+      "grad_norm": 9.213792271740255,
+      "learning_rate": 5.348760330578512e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -318.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.71875,
+      "step": 827
+    },
+    {
+      "epoch": 1.0159509202453987,
+      "grad_norm": 12.070002953910205,
+      "learning_rate": 5.345454545454545e-07,
+      "logits/chosen": -0.025634765625,
+      "logits/rejected": -0.220703125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -310.0,
+      "loss": 0.0745,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.7265625,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.6875,
+      "step": 828
+    },
+    {
+      "epoch": 1.0171779141104293,
+      "grad_norm": 14.6095484944048,
+      "learning_rate": 5.342148760330579e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0873,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 2.0625,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.03125,
+      "step": 829
+    },
+    {
+      "epoch": 1.01840490797546,
+      "grad_norm": 10.563003784443062,
+      "learning_rate": 5.338842975206612e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0528,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 2.578125,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -6.625,
+      "step": 830
+    },
+    {
+      "epoch": 1.019631901840491,
+      "grad_norm": 6.950594992951885,
+      "learning_rate": 5.335537190082644e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0425,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.328125,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -6.75,
+      "step": 831
+    },
+    {
+      "epoch": 1.0208588957055216,
+      "grad_norm": 18.903093292372017,
+      "learning_rate": 5.332231404958677e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0753,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -6.1875,
+      "step": 832
+    },
+    {
+      "epoch": 1.0220858895705522,
+      "grad_norm": 20.846144434558045,
+      "learning_rate": 5.32892561983471e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -442.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0754,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -6.625,
+      "step": 833
+    },
+    {
+      "epoch": 1.023312883435583,
+      "grad_norm": 8.267162808665637,
+      "learning_rate": 5.325619834710743e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.173828125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0429,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -6.25,
+      "step": 834
+    },
+    {
+      "epoch": 1.0245398773006136,
+      "grad_norm": 10.37076437566513,
+      "learning_rate": 5.322314049586777e-07,
+      "logits/chosen": -0.0859375,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0656,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -6.3125,
+      "step": 835
+    },
+    {
+      "epoch": 1.0257668711656442,
+      "grad_norm": 15.10396868081453,
+      "learning_rate": 5.31900826446281e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0572,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -6.9375,
+      "step": 836
+    },
+    {
+      "epoch": 1.026993865030675,
+      "grad_norm": 9.458003649126487,
+      "learning_rate": 5.315702479338843e-07,
+      "logits/chosen": -0.041015625,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0542,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 2.0625,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -6.75,
+      "step": 837
+    },
+    {
+      "epoch": 1.0282208588957056,
+      "grad_norm": 9.28143966396222,
+      "learning_rate": 5.312396694214876e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0463,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -7.5,
+      "step": 838
+    },
+    {
+      "epoch": 1.0294478527607362,
+      "grad_norm": 26.604715129647236,
+      "learning_rate": 5.309090909090909e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0866,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -6.84375,
+      "step": 839
+    },
+    {
+      "epoch": 1.030674846625767,
+      "grad_norm": 9.15851027720581,
+      "learning_rate": 5.305785123966941e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0476,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -7.84375,
+      "step": 840
+    },
+    {
+      "epoch": 1.0319018404907976,
+      "grad_norm": 9.875999818863187,
+      "learning_rate": 5.302479338842975e-07,
+      "logits/chosen": -0.0849609375,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0536,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.46875,
+      "step": 841
+    },
+    {
+      "epoch": 1.0331288343558283,
+      "grad_norm": 8.966946429222945,
+      "learning_rate": 5.299173553719008e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -448.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0535,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -6.96875,
+      "step": 842
+    },
+    {
+      "epoch": 1.034355828220859,
+      "grad_norm": 8.533895413196568,
+      "learning_rate": 5.295867768595041e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -450.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.0625,
+      "step": 843
+    },
+    {
+      "epoch": 1.0355828220858896,
+      "grad_norm": 8.091996676249778,
+      "learning_rate": 5.292561983471074e-07,
+      "logits/chosen": -0.07861328125,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0398,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.03125,
+      "step": 844
+    },
+    {
+      "epoch": 1.0368098159509203,
+      "grad_norm": 12.849424838058209,
+      "learning_rate": 5.289256198347107e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0503,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -6.28125,
+      "step": 845
+    },
+    {
+      "epoch": 1.038036809815951,
+      "grad_norm": 8.52014381356995,
+      "learning_rate": 5.285950413223141e-07,
+      "logits/chosen": -0.0546875,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0457,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.1875,
+      "step": 846
+    },
+    {
+      "epoch": 1.0392638036809816,
+      "grad_norm": 9.894794847731864,
+      "learning_rate": 5.282644628099174e-07,
+      "logits/chosen": -0.07861328125,
+      "logits/rejected": -0.2353515625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.6875,
+      "step": 847
+    },
+    {
+      "epoch": 1.0404907975460123,
+      "grad_norm": 7.67098466898408,
+      "learning_rate": 5.279338842975206e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0467,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.8515625,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.5625,
+      "step": 848
+    },
+    {
+      "epoch": 1.041717791411043,
+      "grad_norm": 13.659276587750096,
+      "learning_rate": 5.276033057851239e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.2080078125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0674,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.0,
+      "step": 849
+    },
+    {
+      "epoch": 1.0429447852760736,
+      "grad_norm": 8.672069845350178,
+      "learning_rate": 5.272727272727272e-07,
+      "logits/chosen": -0.1513671875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -342.0,
+      "loss": 0.032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.34375,
+      "step": 850
+    },
+    {
+      "epoch": 1.0441717791411043,
+      "grad_norm": 4.966708191691141,
+      "learning_rate": 5.269421487603305e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0252,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -7.5625,
+      "step": 851
+    },
+    {
+      "epoch": 1.045398773006135,
+      "grad_norm": 7.342481591915201,
+      "learning_rate": 5.266115702479339e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0379,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.828125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -6.90625,
+      "step": 852
+    },
+    {
+      "epoch": 1.0466257668711656,
+      "grad_norm": 8.819410735961545,
+      "learning_rate": 5.262809917355372e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.9375,
+      "step": 853
+    },
+    {
+      "epoch": 1.0478527607361963,
+      "grad_norm": 13.464418269579307,
+      "learning_rate": 5.259504132231405e-07,
+      "logits/chosen": -0.0966796875,
+      "logits/rejected": -0.232421875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0633,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 7.65625,
+      "rewards/rejected": -6.71875,
+      "step": 854
+    },
+    {
+      "epoch": 1.049079754601227,
+      "grad_norm": 12.329495080989613,
+      "learning_rate": 5.256198347107438e-07,
+      "logits/chosen": -0.038818359375,
+      "logits/rejected": -0.185546875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0646,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.84375,
+      "step": 855
+    },
+    {
+      "epoch": 1.0503067484662576,
+      "grad_norm": 13.38637868578377,
+      "learning_rate": 5.252892561983471e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0713,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.09375,
+      "step": 856
+    },
+    {
+      "epoch": 1.0515337423312883,
+      "grad_norm": 9.793442909461401,
+      "learning_rate": 5.249586776859503e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -364.0,
+      "loss": 0.047,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -7.625,
+      "step": 857
+    },
+    {
+      "epoch": 1.052760736196319,
+      "grad_norm": 6.67438475020692,
+      "learning_rate": 5.246280991735537e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.1904296875,
+      "logps/chosen": -318.0,
+      "logps/rejected": -304.0,
+      "loss": 0.0366,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.84375,
+      "step": 858
+    },
+    {
+      "epoch": 1.0539877300613496,
+      "grad_norm": 8.529881979528566,
+      "learning_rate": 5.24297520661157e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0383,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.76953125,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.84375,
+      "step": 859
+    },
+    {
+      "epoch": 1.0552147239263803,
+      "grad_norm": 11.085862425760341,
+      "learning_rate": 5.239669421487603e-07,
+      "logits/chosen": -0.0791015625,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0715,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.46875,
+      "rewards/margins": 7.46875,
+      "rewards/rejected": -7.0,
+      "step": 860
+    },
+    {
+      "epoch": 1.056441717791411,
+      "grad_norm": 9.866985209297734,
+      "learning_rate": 5.236363636363636e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0442,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4296875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -7.59375,
+      "step": 861
+    },
+    {
+      "epoch": 1.0576687116564418,
+      "grad_norm": 9.885391182604536,
+      "learning_rate": 5.233057851239669e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0428,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.36328125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.375,
+      "step": 862
+    },
+    {
+      "epoch": 1.0588957055214725,
+      "grad_norm": 5.796756692523751,
+      "learning_rate": 5.229752066115702e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0241,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.953125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -7.9375,
+      "step": 863
+    },
+    {
+      "epoch": 1.0601226993865032,
+      "grad_norm": 9.27987197358265,
+      "learning_rate": 5.226446280991736e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -368.0,
+      "loss": 0.048,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.5625,
+      "step": 864
+    },
+    {
+      "epoch": 1.0613496932515338,
+      "grad_norm": 8.3395730894386,
+      "learning_rate": 5.223140495867769e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0338,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.859375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.8125,
+      "step": 865
+    },
+    {
+      "epoch": 1.0625766871165645,
+      "grad_norm": 6.886729358178275,
+      "learning_rate": 5.219834710743801e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -348.0,
+      "loss": 0.032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.6875,
+      "step": 866
+    },
+    {
+      "epoch": 1.0638036809815952,
+      "grad_norm": 16.605968137616845,
+      "learning_rate": 5.216528925619834e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0846,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.8125,
+      "step": 867
+    },
+    {
+      "epoch": 1.0650306748466258,
+      "grad_norm": 9.468389714978352,
+      "learning_rate": 5.213223140495867e-07,
+      "logits/chosen": 0.00176239013671875,
+      "logits/rejected": -0.1982421875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -308.0,
+      "loss": 0.0545,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.73046875,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -7.34375,
+      "step": 868
+    },
+    {
+      "epoch": 1.0662576687116565,
+      "grad_norm": 7.444446735409014,
+      "learning_rate": 5.2099173553719e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0385,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.1875,
+      "step": 869
+    },
+    {
+      "epoch": 1.0674846625766872,
+      "grad_norm": 9.293616058876756,
+      "learning_rate": 5.206611570247934e-07,
+      "logits/chosen": -0.06591796875,
+      "logits/rejected": -0.2470703125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0503,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.78125,
+      "step": 870
+    },
+    {
+      "epoch": 1.0687116564417178,
+      "grad_norm": 14.736932461238908,
+      "learning_rate": 5.203305785123967e-07,
+      "logits/chosen": -0.0703125,
+      "logits/rejected": -0.2353515625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0651,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.83984375,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.15625,
+      "step": 871
+    },
+    {
+      "epoch": 1.0699386503067485,
+      "grad_norm": 10.350348873115376,
+      "learning_rate": 5.2e-07,
+      "logits/chosen": -0.087890625,
+      "logits/rejected": -0.2177734375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -318.0,
+      "loss": 0.0463,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.5625,
+      "step": 872
+    },
+    {
+      "epoch": 1.0711656441717792,
+      "grad_norm": 10.092656085910821,
+      "learning_rate": 5.196694214876033e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0513,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.125,
+      "step": 873
+    },
+    {
+      "epoch": 1.0723926380368098,
+      "grad_norm": 18.614352522823648,
+      "learning_rate": 5.193388429752066e-07,
+      "logits/chosen": -0.2001953125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0732,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.71875,
+      "step": 874
+    },
+    {
+      "epoch": 1.0736196319018405,
+      "grad_norm": 7.961276364458803,
+      "learning_rate": 5.190082644628098e-07,
+      "logits/chosen": -0.08203125,
+      "logits/rejected": -0.24609375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0406,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.71875,
+      "step": 875
+    },
+    {
+      "epoch": 1.0748466257668712,
+      "grad_norm": 10.357878513050897,
+      "learning_rate": 5.186776859504132e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0483,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.375,
+      "step": 876
+    },
+    {
+      "epoch": 1.0760736196319018,
+      "grad_norm": 10.744500955796752,
+      "learning_rate": 5.183471074380165e-07,
+      "logits/chosen": -0.0302734375,
+      "logits/rejected": -0.1826171875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.427734375,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.59375,
+      "step": 877
+    },
+    {
+      "epoch": 1.0773006134969325,
+      "grad_norm": 12.618829108696058,
+      "learning_rate": 5.180165289256198e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0761,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.376953125,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -7.75,
+      "step": 878
+    },
+    {
+      "epoch": 1.0785276073619632,
+      "grad_norm": 8.80156369716181,
+      "learning_rate": 5.176859504132231e-07,
+      "logits/chosen": -0.06982421875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0375,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.65625,
+      "step": 879
+    },
+    {
+      "epoch": 1.0797546012269938,
+      "grad_norm": 11.402516366286292,
+      "learning_rate": 5.173553719008264e-07,
+      "logits/chosen": -0.07861328125,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0504,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.062255859375,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.75,
+      "step": 880
+    },
+    {
+      "epoch": 1.0809815950920245,
+      "grad_norm": 11.031397024556938,
+      "learning_rate": 5.170247933884298e-07,
+      "logits/chosen": -0.048828125,
+      "logits/rejected": -0.1884765625,
+      "logps/chosen": -348.0,
+      "logps/rejected": -302.0,
+      "loss": 0.0618,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -6.8125,
+      "step": 881
+    },
+    {
+      "epoch": 1.0822085889570552,
+      "grad_norm": 18.547130207267188,
+      "learning_rate": 5.166942148760331e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0755,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.125,
+      "step": 882
+    },
+    {
+      "epoch": 1.0834355828220859,
+      "grad_norm": 6.6527500793888805,
+      "learning_rate": 5.163636363636364e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0289,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -6.6875,
+      "step": 883
+    },
+    {
+      "epoch": 1.0846625766871165,
+      "grad_norm": 6.861231219217128,
+      "learning_rate": 5.160330578512396e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -426.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0301,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.171875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -7.5,
+      "step": 884
+    },
+    {
+      "epoch": 1.0858895705521472,
+      "grad_norm": 9.381597515049744,
+      "learning_rate": 5.157024793388429e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0562,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.78125,
+      "step": 885
+    },
+    {
+      "epoch": 1.0871165644171779,
+      "grad_norm": 8.908837796346583,
+      "learning_rate": 5.153719008264462e-07,
+      "logits/chosen": -0.041748046875,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0518,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 7.75,
+      "rewards/rejected": -6.46875,
+      "step": 886
+    },
+    {
+      "epoch": 1.0883435582822085,
+      "grad_norm": 11.48855243783124,
+      "learning_rate": 5.150413223140496e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0543,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.75,
+      "step": 887
+    },
+    {
+      "epoch": 1.0895705521472392,
+      "grad_norm": 6.793415855369003,
+      "learning_rate": 5.147107438016529e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0336,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.34375,
+      "step": 888
+    },
+    {
+      "epoch": 1.0907975460122699,
+      "grad_norm": 15.354710995482474,
+      "learning_rate": 5.143801652892562e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -450.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0741,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.0625,
+      "step": 889
+    },
+    {
+      "epoch": 1.0920245398773005,
+      "grad_norm": 9.766896381796215,
+      "learning_rate": 5.140495867768595e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0492,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -7.0,
+      "step": 890
+    },
+    {
+      "epoch": 1.0932515337423312,
+      "grad_norm": 9.533700281126114,
+      "learning_rate": 5.137190082644628e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0521,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.8125,
+      "step": 891
+    },
+    {
+      "epoch": 1.0944785276073619,
+      "grad_norm": 8.94436680718413,
+      "learning_rate": 5.133884297520661e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -340.0,
+      "loss": 0.049,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.875,
+      "step": 892
+    },
+    {
+      "epoch": 1.0957055214723925,
+      "grad_norm": 13.394144993964538,
+      "learning_rate": 5.130578512396694e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.2099609375,
+      "logps/chosen": -316.0,
+      "logps/rejected": -312.0,
+      "loss": 0.0707,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 0.8125,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.1875,
+      "step": 893
+    },
+    {
+      "epoch": 1.0969325153374232,
+      "grad_norm": 17.301836509277987,
+      "learning_rate": 5.127272727272727e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0535,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.7265625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.125,
+      "step": 894
+    },
+    {
+      "epoch": 1.098159509202454,
+      "grad_norm": 13.580959238382313,
+      "learning_rate": 5.12396694214876e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0519,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -7.03125,
+      "step": 895
+    },
+    {
+      "epoch": 1.0993865030674848,
+      "grad_norm": 10.698325211425429,
+      "learning_rate": 5.120661157024793e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0481,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.46875,
+      "step": 896
+    },
+    {
+      "epoch": 1.1006134969325154,
+      "grad_norm": 11.842993344318062,
+      "learning_rate": 5.117355371900826e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -440.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0496,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.84375,
+      "step": 897
+    },
+    {
+      "epoch": 1.101840490797546,
+      "grad_norm": 7.993601971885361,
+      "learning_rate": 5.114049586776859e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0393,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.0625,
+      "step": 898
+    },
+    {
+      "epoch": 1.1030674846625768,
+      "grad_norm": 7.485760340921191,
+      "learning_rate": 5.110743801652893e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0429,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.287109375,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -8.0625,
+      "step": 899
+    },
+    {
+      "epoch": 1.1042944785276074,
+      "grad_norm": 10.22220098269188,
+      "learning_rate": 5.107438016528926e-07,
+      "logits/chosen": -0.10205078125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -380.0,
+      "loss": 0.046,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.80859375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.875,
+      "step": 900
+    },
+    {
+      "epoch": 1.105521472392638,
+      "grad_norm": 11.370160710024297,
+      "learning_rate": 5.104132231404959e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -490.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0534,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.828125,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -8.6875,
+      "step": 901
+    },
+    {
+      "epoch": 1.1067484662576688,
+      "grad_norm": 12.269173865343777,
+      "learning_rate": 5.100826446280991e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0518,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.91015625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.5625,
+      "step": 902
+    },
+    {
+      "epoch": 1.1079754601226994,
+      "grad_norm": 10.521897818355166,
+      "learning_rate": 5.097520661157024e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0463,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.1875,
+      "step": 903
+    },
+    {
+      "epoch": 1.10920245398773,
+      "grad_norm": 6.33098217292226,
+      "learning_rate": 5.094214876033057e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0288,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.5,
+      "step": 904
+    },
+    {
+      "epoch": 1.1104294478527608,
+      "grad_norm": 10.008456465270969,
+      "learning_rate": 5.090909090909091e-07,
+      "logits/chosen": -0.330078125,
+      "logits/rejected": -0.54296875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0393,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -8.4375,
+      "step": 905
+    },
+    {
+      "epoch": 1.1116564417177914,
+      "grad_norm": 9.946359264363899,
+      "learning_rate": 5.087603305785124e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -336.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0528,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.126953125,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -7.90625,
+      "step": 906
+    },
+    {
+      "epoch": 1.112883435582822,
+      "grad_norm": 13.90671388757752,
+      "learning_rate": 5.084297520661157e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0595,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.625,
+      "step": 907
+    },
+    {
+      "epoch": 1.1141104294478528,
+      "grad_norm": 17.130329509275274,
+      "learning_rate": 5.08099173553719e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0457,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.78125,
+      "step": 908
+    },
+    {
+      "epoch": 1.1153374233128834,
+      "grad_norm": 13.27945349361827,
+      "learning_rate": 5.077685950413223e-07,
+      "logits/chosen": -0.11572265625,
+      "logits/rejected": -0.228515625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0579,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.375,
+      "step": 909
+    },
+    {
+      "epoch": 1.116564417177914,
+      "grad_norm": 9.737802290234768,
+      "learning_rate": 5.074380165289256e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0329,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.375,
+      "step": 910
+    },
+    {
+      "epoch": 1.1177914110429448,
+      "grad_norm": 6.685651209591217,
+      "learning_rate": 5.071074380165289e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -466.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0268,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -9.0,
+      "step": 911
+    },
+    {
+      "epoch": 1.1190184049079754,
+      "grad_norm": 13.867209331156516,
+      "learning_rate": 5.067768595041322e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.19140625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0517,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.8359375,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.8125,
+      "step": 912
+    },
+    {
+      "epoch": 1.120245398773006,
+      "grad_norm": 11.283456894779777,
+      "learning_rate": 5.064462809917355e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -322.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0591,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.625,
+      "step": 913
+    },
+    {
+      "epoch": 1.1214723926380368,
+      "grad_norm": 13.891782784573822,
+      "learning_rate": 5.061157024793388e-07,
+      "logits/chosen": -0.09423828125,
+      "logits/rejected": -0.26953125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0699,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.625,
+      "step": 914
+    },
+    {
+      "epoch": 1.1226993865030674,
+      "grad_norm": 15.737683653007828,
+      "learning_rate": 5.057851239669421e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0565,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.0,
+      "step": 915
+    },
+    {
+      "epoch": 1.123926380368098,
+      "grad_norm": 9.209835377946382,
+      "learning_rate": 5.054545454545454e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0438,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.203125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.90625,
+      "step": 916
+    },
+    {
+      "epoch": 1.1251533742331288,
+      "grad_norm": 14.318300360080142,
+      "learning_rate": 5.051239669421488e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -404.0,
+      "loss": 0.072,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.875,
+      "step": 917
+    },
+    {
+      "epoch": 1.1263803680981594,
+      "grad_norm": 16.551514820318832,
+      "learning_rate": 5.047933884297521e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0695,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.28125,
+      "step": 918
+    },
+    {
+      "epoch": 1.1276073619631901,
+      "grad_norm": 11.082245597349267,
+      "learning_rate": 5.044628099173554e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0584,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.90625,
+      "step": 919
+    },
+    {
+      "epoch": 1.1288343558282208,
+      "grad_norm": 12.619871004485745,
+      "learning_rate": 5.041322314049586e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0433,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -7.71875,
+      "step": 920
+    },
+    {
+      "epoch": 1.1300613496932514,
+      "grad_norm": 10.77826696391865,
+      "learning_rate": 5.038016528925619e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0345,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.6484375,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.6875,
+      "step": 921
+    },
+    {
+      "epoch": 1.1312883435582821,
+      "grad_norm": 9.71506841489295,
+      "learning_rate": 5.034710743801652e-07,
+      "logits/chosen": -0.12158203125,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0473,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.9140625,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -6.875,
+      "step": 922
+    },
+    {
+      "epoch": 1.132515337423313,
+      "grad_norm": 8.986837909388852,
+      "learning_rate": 5.031404958677686e-07,
+      "logits/chosen": -0.1064453125,
+      "logits/rejected": -0.2041015625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0394,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 8.3125,
+      "rewards/rejected": -6.71875,
+      "step": 923
+    },
+    {
+      "epoch": 1.1337423312883437,
+      "grad_norm": 11.240488448077638,
+      "learning_rate": 5.028099173553719e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -340.0,
+      "loss": 0.048,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -7.53125,
+      "step": 924
+    },
+    {
+      "epoch": 1.1349693251533743,
+      "grad_norm": 10.37676153526059,
+      "learning_rate": 5.024793388429752e-07,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.5,
+      "logps/chosen": -372.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0385,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.66796875,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -7.90625,
+      "step": 925
+    },
+    {
+      "epoch": 1.136196319018405,
+      "grad_norm": 14.556750345009446,
+      "learning_rate": 5.021487603305785e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0586,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.859375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -7.5625,
+      "step": 926
+    },
+    {
+      "epoch": 1.1374233128834357,
+      "grad_norm": 8.6263295939874,
+      "learning_rate": 5.018181818181818e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0309,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.625,
+      "step": 927
+    },
+    {
+      "epoch": 1.1386503067484663,
+      "grad_norm": 7.620881147688411,
+      "learning_rate": 5.014876033057852e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0414,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.53125,
+      "step": 928
+    },
+    {
+      "epoch": 1.139877300613497,
+      "grad_norm": 11.937716031232593,
+      "learning_rate": 5.011570247933884e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -426.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0443,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.4375,
+      "step": 929
+    },
+    {
+      "epoch": 1.1411042944785277,
+      "grad_norm": 25.46977596486142,
+      "learning_rate": 5.008264462809917e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0556,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.25,
+      "step": 930
+    },
+    {
+      "epoch": 1.1423312883435583,
+      "grad_norm": 12.357953826396704,
+      "learning_rate": 5.00495867768595e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0626,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.7265625,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -7.1875,
+      "step": 931
+    },
+    {
+      "epoch": 1.143558282208589,
+      "grad_norm": 10.124449885714515,
+      "learning_rate": 5.001652892561983e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0487,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.87109375,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.75,
+      "step": 932
+    },
+    {
+      "epoch": 1.1447852760736197,
+      "grad_norm": 10.186523170112832,
+      "learning_rate": 4.998347107438016e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0407,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.53125,
+      "step": 933
+    },
+    {
+      "epoch": 1.1460122699386504,
+      "grad_norm": 14.784919527567176,
+      "learning_rate": 4.99504132231405e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -344.0,
+      "loss": 0.055,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.09375,
+      "step": 934
+    },
+    {
+      "epoch": 1.147239263803681,
+      "grad_norm": 8.084085375609561,
+      "learning_rate": 4.991735537190083e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0339,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.5625,
+      "step": 935
+    },
+    {
+      "epoch": 1.1484662576687117,
+      "grad_norm": 10.090762024483787,
+      "learning_rate": 4.988429752066116e-07,
+      "logits/chosen": -0.06787109375,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0428,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.66796875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.4375,
+      "step": 936
+    },
+    {
+      "epoch": 1.1496932515337424,
+      "grad_norm": 11.391883354259097,
+      "learning_rate": 4.985123966942148e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0446,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.4375,
+      "step": 937
+    },
+    {
+      "epoch": 1.150920245398773,
+      "grad_norm": 17.337934806380964,
+      "learning_rate": 4.981818181818181e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.2001953125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0702,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.73828125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.25,
+      "step": 938
+    },
+    {
+      "epoch": 1.1521472392638037,
+      "grad_norm": 6.607905978535942,
+      "learning_rate": 4.978512396694214e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -444.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0298,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -8.8125,
+      "step": 939
+    },
+    {
+      "epoch": 1.1533742331288344,
+      "grad_norm": 7.07231413487028,
+      "learning_rate": 4.975206611570248e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0283,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.0,
+      "step": 940
+    },
+    {
+      "epoch": 1.154601226993865,
+      "grad_norm": 14.928874081631259,
+      "learning_rate": 4.971900826446281e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0622,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.71875,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -7.8125,
+      "step": 941
+    },
+    {
+      "epoch": 1.1558282208588957,
+      "grad_norm": 17.516711144244283,
+      "learning_rate": 4.968595041322314e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0501,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -9.3125,
+      "step": 942
+    },
+    {
+      "epoch": 1.1570552147239264,
+      "grad_norm": 8.961156359700125,
+      "learning_rate": 4.965289256198347e-07,
+      "logits/chosen": -0.10400390625,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0371,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.62109375,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.0,
+      "step": 943
+    },
+    {
+      "epoch": 1.158282208588957,
+      "grad_norm": 8.12773430831165,
+      "learning_rate": 4.96198347107438e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0336,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.75,
+      "step": 944
+    },
+    {
+      "epoch": 1.1595092024539877,
+      "grad_norm": 11.434978350030235,
+      "learning_rate": 4.958677685950413e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0455,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.0,
+      "step": 945
+    },
+    {
+      "epoch": 1.1607361963190184,
+      "grad_norm": 10.093045663484858,
+      "learning_rate": 4.955371900826446e-07,
+      "logits/chosen": -0.1240234375,
+      "logits/rejected": -0.205078125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0429,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.53125,
+      "step": 946
+    },
+    {
+      "epoch": 1.161963190184049,
+      "grad_norm": 11.765426924290539,
+      "learning_rate": 4.952066115702479e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -342.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0689,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.5,
+      "step": 947
+    },
+    {
+      "epoch": 1.1631901840490797,
+      "grad_norm": 12.622464974038337,
+      "learning_rate": 4.948760330578512e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0432,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -7.46875,
+      "step": 948
+    },
+    {
+      "epoch": 1.1644171779141104,
+      "grad_norm": 11.580415695153775,
+      "learning_rate": 4.945454545454545e-07,
+      "logits/chosen": -0.119140625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0559,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -7.6875,
+      "step": 949
+    },
+    {
+      "epoch": 1.165644171779141,
+      "grad_norm": 9.115099300613052,
+      "learning_rate": 4.942148760330578e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -334.0,
+      "logps/rejected": -310.0,
+      "loss": 0.0411,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.7890625,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.65625,
+      "step": 950
+    },
+    {
+      "epoch": 1.1668711656441717,
+      "grad_norm": 11.047710529408716,
+      "learning_rate": 4.938842975206611e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0529,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.73046875,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -7.40625,
+      "step": 951
+    },
+    {
+      "epoch": 1.1680981595092024,
+      "grad_norm": 13.283612267053757,
+      "learning_rate": 4.935537190082645e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0633,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.84375,
+      "step": 952
+    },
+    {
+      "epoch": 1.169325153374233,
+      "grad_norm": 12.688088508685484,
+      "learning_rate": 4.932231404958678e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0515,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -7.0625,
+      "step": 953
+    },
+    {
+      "epoch": 1.1705521472392637,
+      "grad_norm": 13.72625399518082,
+      "learning_rate": 4.928925619834711e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -316.0,
+      "logps/rejected": -294.0,
+      "loss": 0.0694,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.17578125,
+      "rewards/margins": 7.84375,
+      "rewards/rejected": -7.6875,
+      "step": 954
+    },
+    {
+      "epoch": 1.1717791411042944,
+      "grad_norm": 15.264557868135348,
+      "learning_rate": 4.925619834710743e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.2080078125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0505,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4296875,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.25,
+      "step": 955
+    },
+    {
+      "epoch": 1.173006134969325,
+      "grad_norm": 8.222673231264135,
+      "learning_rate": 4.922314049586776e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -310.0,
+      "logps/rejected": -320.0,
+      "loss": 0.0373,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 7.875,
+      "rewards/rejected": -7.125,
+      "step": 956
+    },
+    {
+      "epoch": 1.1742331288343557,
+      "grad_norm": 8.70101877312667,
+      "learning_rate": 4.919008264462809e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0379,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.34375,
+      "step": 957
+    },
+    {
+      "epoch": 1.1754601226993864,
+      "grad_norm": 8.625233525867843,
+      "learning_rate": 4.915702479338843e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0329,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.0625,
+      "step": 958
+    },
+    {
+      "epoch": 1.1766871165644173,
+      "grad_norm": 8.831414998360419,
+      "learning_rate": 4.912396694214876e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0393,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.71875,
+      "step": 959
+    },
+    {
+      "epoch": 1.177914110429448,
+      "grad_norm": 7.156930699104763,
+      "learning_rate": 4.909090909090909e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0348,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.265625,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -8.25,
+      "step": 960
+    },
+    {
+      "epoch": 1.1791411042944786,
+      "grad_norm": 7.316598874843914,
+      "learning_rate": 4.905785123966942e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -356.0,
+      "loss": 0.032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.125,
+      "step": 961
+    },
+    {
+      "epoch": 1.1803680981595093,
+      "grad_norm": 15.127473084571868,
+      "learning_rate": 4.902479338842975e-07,
+      "logits/chosen": -0.1220703125,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0505,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.96875,
+      "step": 962
+    },
+    {
+      "epoch": 1.18159509202454,
+      "grad_norm": 7.350557070973821,
+      "learning_rate": 4.899173553719008e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0355,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -7.90625,
+      "step": 963
+    },
+    {
+      "epoch": 1.1828220858895706,
+      "grad_norm": 7.9216285170357805,
+      "learning_rate": 4.895867768595041e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0269,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.25,
+      "step": 964
+    },
+    {
+      "epoch": 1.1840490797546013,
+      "grad_norm": 17.652292078469603,
+      "learning_rate": 4.892561983471074e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0483,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.55078125,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.5,
+      "step": 965
+    },
+    {
+      "epoch": 1.185276073619632,
+      "grad_norm": 9.347424143338342,
+      "learning_rate": 4.889256198347107e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0449,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.83203125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.1875,
+      "step": 966
+    },
+    {
+      "epoch": 1.1865030674846626,
+      "grad_norm": 8.997313851138234,
+      "learning_rate": 4.88595041322314e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0435,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.71875,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.0,
+      "step": 967
+    },
+    {
+      "epoch": 1.1877300613496933,
+      "grad_norm": 8.350460284446099,
+      "learning_rate": 4.882644628099173e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0278,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -8.9375,
+      "step": 968
+    },
+    {
+      "epoch": 1.188957055214724,
+      "grad_norm": 17.12432481280222,
+      "learning_rate": 4.879338842975206e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -396.0,
+      "loss": 0.063,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.453125,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -8.6875,
+      "step": 969
+    },
+    {
+      "epoch": 1.1901840490797546,
+      "grad_norm": 11.73493140010646,
+      "learning_rate": 4.87603305785124e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -484.0,
+      "loss": 0.0539,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 22.0,
+      "rewards/rejected": -21.25,
+      "step": 970
+    },
+    {
+      "epoch": 1.1914110429447853,
+      "grad_norm": 11.251800364854649,
+      "learning_rate": 4.872727272727273e-07,
+      "logits/chosen": -0.1357421875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0577,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.50390625,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.4375,
+      "step": 971
+    },
+    {
+      "epoch": 1.192638036809816,
+      "grad_norm": 12.538848815995712,
+      "learning_rate": 4.869421487603306e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -338.0,
+      "logps/rejected": -292.0,
+      "loss": 0.0648,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1181640625,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -7.875,
+      "step": 972
+    },
+    {
+      "epoch": 1.1938650306748466,
+      "grad_norm": 8.47785010044017,
+      "learning_rate": 4.866115702479338e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0277,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.5,
+      "step": 973
+    },
+    {
+      "epoch": 1.1950920245398773,
+      "grad_norm": 9.052453334123884,
+      "learning_rate": 4.862809917355371e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0343,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -8.875,
+      "step": 974
+    },
+    {
+      "epoch": 1.196319018404908,
+      "grad_norm": 12.01099472634397,
+      "learning_rate": 4.859504132231405e-07,
+      "logits/chosen": -0.095703125,
+      "logits/rejected": -0.22265625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -314.0,
+      "loss": 0.041,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.78125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -7.96875,
+      "step": 975
+    },
+    {
+      "epoch": 1.1975460122699386,
+      "grad_norm": 11.988662613154055,
+      "learning_rate": 4.856198347107438e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0538,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.375,
+      "step": 976
+    },
+    {
+      "epoch": 1.1987730061349693,
+      "grad_norm": 11.13765858997807,
+      "learning_rate": 4.852892561983471e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0288,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -8.8125,
+      "step": 977
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 13.793986032637335,
+      "learning_rate": 4.849586776859504e-07,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0739,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.3125,
+      "step": 978
+    },
+    {
+      "epoch": 1.2012269938650306,
+      "grad_norm": 8.480774562673568,
+      "learning_rate": 4.846280991735537e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0307,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.4375,
+      "step": 979
+    },
+    {
+      "epoch": 1.2024539877300613,
+      "grad_norm": 4.365430571481909,
+      "learning_rate": 4.84297520661157e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0198,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.0625,
+      "step": 980
+    },
+    {
+      "epoch": 1.203680981595092,
+      "grad_norm": 13.729307845818116,
+      "learning_rate": 4.839669421487604e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0469,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.5625,
+      "step": 981
+    },
+    {
+      "epoch": 1.2049079754601226,
+      "grad_norm": 10.207687106718279,
+      "learning_rate": 4.836363636363636e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0494,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.875,
+      "step": 982
+    },
+    {
+      "epoch": 1.2061349693251533,
+      "grad_norm": 17.058047206794242,
+      "learning_rate": 4.833057851239669e-07,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0847,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -6.8125,
+      "step": 983
+    },
+    {
+      "epoch": 1.207361963190184,
+      "grad_norm": 12.189727136679215,
+      "learning_rate": 4.829752066115702e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0584,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.75,
+      "step": 984
+    },
+    {
+      "epoch": 1.2085889570552146,
+      "grad_norm": 12.801014204687064,
+      "learning_rate": 4.826446280991735e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -344.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0543,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.46875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.375,
+      "step": 985
+    },
+    {
+      "epoch": 1.2098159509202455,
+      "grad_norm": 9.609094722284011,
+      "learning_rate": 4.823140495867768e-07,
+      "logits/chosen": -0.050537109375,
+      "logits/rejected": -0.1962890625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0361,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -6.875,
+      "step": 986
+    },
+    {
+      "epoch": 1.2110429447852762,
+      "grad_norm": 13.218029076540835,
+      "learning_rate": 4.819834710743802e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -452.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0483,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -7.78125,
+      "step": 987
+    },
+    {
+      "epoch": 1.2122699386503069,
+      "grad_norm": 8.141541687744194,
+      "learning_rate": 4.816528925619835e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0309,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.9921875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.9375,
+      "step": 988
+    },
+    {
+      "epoch": 1.2134969325153375,
+      "grad_norm": 12.298643628948078,
+      "learning_rate": 4.813223140495868e-07,
+      "logits/chosen": -0.107421875,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -318.0,
+      "loss": 0.0631,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.67578125,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -7.40625,
+      "step": 989
+    },
+    {
+      "epoch": 1.2147239263803682,
+      "grad_norm": 5.79863996841087,
+      "learning_rate": 4.809917355371901e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0325,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.625,
+      "step": 990
+    },
+    {
+      "epoch": 1.2159509202453989,
+      "grad_norm": 6.760236470696033,
+      "learning_rate": 4.806611570247933e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0284,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.0625,
+      "step": 991
+    },
+    {
+      "epoch": 1.2171779141104295,
+      "grad_norm": 10.672410560638603,
+      "learning_rate": 4.803305785123966e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0472,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.5625,
+      "step": 992
+    },
+    {
+      "epoch": 1.2184049079754602,
+      "grad_norm": 9.716010918159654,
+      "learning_rate": 4.8e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0385,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.96875,
+      "step": 993
+    },
+    {
+      "epoch": 1.2196319018404909,
+      "grad_norm": 10.42174132730132,
+      "learning_rate": 4.796694214876033e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0418,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.625,
+      "step": 994
+    },
+    {
+      "epoch": 1.2208588957055215,
+      "grad_norm": 11.123972657771333,
+      "learning_rate": 4.793388429752066e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0373,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.462890625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.75,
+      "step": 995
+    },
+    {
+      "epoch": 1.2220858895705522,
+      "grad_norm": 25.75709421241266,
+      "learning_rate": 4.790082644628099e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0693,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.41015625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.4375,
+      "step": 996
+    },
+    {
+      "epoch": 1.2233128834355829,
+      "grad_norm": 9.409931474479396,
+      "learning_rate": 4.786776859504132e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0363,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.0,
+      "step": 997
+    },
+    {
+      "epoch": 1.2245398773006135,
+      "grad_norm": 9.490599131683323,
+      "learning_rate": 4.783471074380165e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -442.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0311,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.125,
+      "step": 998
+    },
+    {
+      "epoch": 1.2257668711656442,
+      "grad_norm": 8.127599963420865,
+      "learning_rate": 4.780165289256199e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0319,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.25,
+      "step": 999
+    },
+    {
+      "epoch": 1.2269938650306749,
+      "grad_norm": 12.601959839650176,
+      "learning_rate": 4.776859504132231e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0659,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -7.8125,
+      "step": 1000
+    },
+    {
+      "epoch": 1.2282208588957055,
+      "grad_norm": 10.056271597433025,
+      "learning_rate": 4.773553719008264e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0356,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.3125,
+      "step": 1001
+    },
+    {
+      "epoch": 1.2294478527607362,
+      "grad_norm": 17.964616987432837,
+      "learning_rate": 4.770247933884297e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0573,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -7.96875,
+      "step": 1002
+    },
+    {
+      "epoch": 1.2306748466257669,
+      "grad_norm": 10.775894884702035,
+      "learning_rate": 4.7669421487603303e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0485,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.92578125,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.25,
+      "step": 1003
+    },
+    {
+      "epoch": 1.2319018404907975,
+      "grad_norm": 8.116589015572577,
+      "learning_rate": 4.7636363636363635e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -308.0,
+      "loss": 0.0375,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.375,
+      "step": 1004
+    },
+    {
+      "epoch": 1.2331288343558282,
+      "grad_norm": 10.13931784187266,
+      "learning_rate": 4.7603305785123966e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0502,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.25,
+      "step": 1005
+    },
+    {
+      "epoch": 1.2343558282208589,
+      "grad_norm": 8.80736282586411,
+      "learning_rate": 4.75702479338843e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0403,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.375,
+      "step": 1006
+    },
+    {
+      "epoch": 1.2355828220858895,
+      "grad_norm": 9.046118916104152,
+      "learning_rate": 4.753719008264463e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0411,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.3125,
+      "step": 1007
+    },
+    {
+      "epoch": 1.2368098159509202,
+      "grad_norm": 12.220589771919323,
+      "learning_rate": 4.750413223140496e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0435,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.59375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.0625,
+      "step": 1008
+    },
+    {
+      "epoch": 1.2380368098159509,
+      "grad_norm": 14.688018091941988,
+      "learning_rate": 4.7471074380165283e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0554,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3828125,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -7.8125,
+      "step": 1009
+    },
+    {
+      "epoch": 1.2392638036809815,
+      "grad_norm": 11.98135604057609,
+      "learning_rate": 4.7438016528925615e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0605,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.032470703125,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.96875,
+      "step": 1010
+    },
+    {
+      "epoch": 1.2404907975460122,
+      "grad_norm": 11.577259622901868,
+      "learning_rate": 4.7404958677685946e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0444,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.64453125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.3125,
+      "step": 1011
+    },
+    {
+      "epoch": 1.2417177914110429,
+      "grad_norm": 16.548853860409746,
+      "learning_rate": 4.737190082644628e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0828,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.8359375,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.0,
+      "step": 1012
+    },
+    {
+      "epoch": 1.2429447852760735,
+      "grad_norm": 8.191313021502108,
+      "learning_rate": 4.733884297520661e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.2265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0457,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.37890625,
+      "rewards/margins": 7.875,
+      "rewards/rejected": -7.5,
+      "step": 1013
+    },
+    {
+      "epoch": 1.2441717791411042,
+      "grad_norm": 10.841806442484222,
+      "learning_rate": 4.730578512396694e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.388671875,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -8.0625,
+      "step": 1014
+    },
+    {
+      "epoch": 1.2453987730061349,
+      "grad_norm": 12.095214544652428,
+      "learning_rate": 4.7272727272727273e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.0625,
+      "step": 1015
+    },
+    {
+      "epoch": 1.2466257668711656,
+      "grad_norm": 20.048615612397384,
+      "learning_rate": 4.7239669421487605e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0901,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.44921875,
+      "rewards/margins": 7.875,
+      "rewards/rejected": -7.4375,
+      "step": 1016
+    },
+    {
+      "epoch": 1.2478527607361962,
+      "grad_norm": 4.909314153216152,
+      "learning_rate": 4.7206611570247937e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0205,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.125,
+      "step": 1017
+    },
+    {
+      "epoch": 1.2490797546012269,
+      "grad_norm": 15.913032240180387,
+      "learning_rate": 4.717355371900826e-07,
+      "logits/chosen": -0.248046875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0537,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.46484375,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.1875,
+      "step": 1018
+    },
+    {
+      "epoch": 1.2503067484662576,
+      "grad_norm": 7.998330769581237,
+      "learning_rate": 4.714049586776859e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0435,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7421875,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.8125,
+      "step": 1019
+    },
+    {
+      "epoch": 1.2515337423312882,
+      "grad_norm": 8.154520742857182,
+      "learning_rate": 4.710743801652892e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0424,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.72265625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.375,
+      "step": 1020
+    },
+    {
+      "epoch": 1.252760736196319,
+      "grad_norm": 10.190607926042517,
+      "learning_rate": 4.7074380165289254e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0457,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.216796875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.8125,
+      "step": 1021
+    },
+    {
+      "epoch": 1.2539877300613496,
+      "grad_norm": 9.551253572647711,
+      "learning_rate": 4.7041322314049585e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0398,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.76171875,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.1875,
+      "step": 1022
+    },
+    {
+      "epoch": 1.2552147239263804,
+      "grad_norm": 9.671012064033086,
+      "learning_rate": 4.7008264462809917e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0413,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -8.875,
+      "step": 1023
+    },
+    {
+      "epoch": 1.2564417177914111,
+      "grad_norm": 12.47562958036183,
+      "learning_rate": 4.697520661157025e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0505,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.5,
+      "step": 1024
+    },
+    {
+      "epoch": 1.2576687116564418,
+      "grad_norm": 12.875244309972366,
+      "learning_rate": 4.694214876033058e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0622,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -8.6875,
+      "step": 1025
+    },
+    {
+      "epoch": 1.2588957055214725,
+      "grad_norm": 9.287273683982628,
+      "learning_rate": 4.69090909090909e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -362.0,
+      "loss": 0.047,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.361328125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.125,
+      "step": 1026
+    },
+    {
+      "epoch": 1.2601226993865031,
+      "grad_norm": 8.982752724660957,
+      "learning_rate": 4.6876033057851234e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -446.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0343,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.78515625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.375,
+      "step": 1027
+    },
+    {
+      "epoch": 1.2613496932515338,
+      "grad_norm": 16.331351941268387,
+      "learning_rate": 4.6842975206611565e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0478,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -9.375,
+      "step": 1028
+    },
+    {
+      "epoch": 1.2625766871165645,
+      "grad_norm": 14.883928893155497,
+      "learning_rate": 4.6809917355371897e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0462,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.5,
+      "step": 1029
+    },
+    {
+      "epoch": 1.2638036809815951,
+      "grad_norm": 8.911098912817012,
+      "learning_rate": 4.677685950413223e-07,
+      "logits/chosen": -0.38671875,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0305,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.4140625,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -9.25,
+      "step": 1030
+    },
+    {
+      "epoch": 1.2650306748466258,
+      "grad_norm": 11.44985669629329,
+      "learning_rate": 4.674380165289256e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0695,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.8359375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.5,
+      "step": 1031
+    },
+    {
+      "epoch": 1.2662576687116565,
+      "grad_norm": 6.783926726533355,
+      "learning_rate": 4.671074380165289e-07,
+      "logits/chosen": -0.07470703125,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0295,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.4375,
+      "step": 1032
+    },
+    {
+      "epoch": 1.2674846625766871,
+      "grad_norm": 13.190345224116596,
+      "learning_rate": 4.6677685950413224e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -430.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0611,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.78515625,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.1875,
+      "step": 1033
+    },
+    {
+      "epoch": 1.2687116564417178,
+      "grad_norm": 6.185510880413833,
+      "learning_rate": 4.6644628099173556e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -460.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0256,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -8.8125,
+      "step": 1034
+    },
+    {
+      "epoch": 1.2699386503067485,
+      "grad_norm": 7.619537783501127,
+      "learning_rate": 4.6611570247933877e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0297,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.625,
+      "step": 1035
+    },
+    {
+      "epoch": 1.2711656441717791,
+      "grad_norm": 13.624124945358417,
+      "learning_rate": 4.657851239669421e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0562,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.65625,
+      "step": 1036
+    },
+    {
+      "epoch": 1.2723926380368098,
+      "grad_norm": 7.586885211554758,
+      "learning_rate": 4.654545454545454e-07,
+      "logits/chosen": -0.35546875,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0282,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.96875,
+      "step": 1037
+    },
+    {
+      "epoch": 1.2736196319018405,
+      "grad_norm": 9.803215310550355,
+      "learning_rate": 4.651239669421487e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0361,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.5,
+      "step": 1038
+    },
+    {
+      "epoch": 1.2748466257668711,
+      "grad_norm": 14.589804231321581,
+      "learning_rate": 4.6479338842975204e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0627,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.8359375,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.59375,
+      "step": 1039
+    },
+    {
+      "epoch": 1.2760736196319018,
+      "grad_norm": 13.800642679130199,
+      "learning_rate": 4.6446280991735536e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0433,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.46875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.375,
+      "step": 1040
+    },
+    {
+      "epoch": 1.2773006134969325,
+      "grad_norm": 6.6360914407229785,
+      "learning_rate": 4.641322314049587e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0192,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.98828125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.125,
+      "step": 1041
+    },
+    {
+      "epoch": 1.2785276073619631,
+      "grad_norm": 10.335577733951094,
+      "learning_rate": 4.63801652892562e-07,
+      "logits/chosen": -0.369140625,
+      "logits/rejected": -0.57421875,
+      "logps/chosen": -444.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0391,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.6875,
+      "step": 1042
+    },
+    {
+      "epoch": 1.2797546012269938,
+      "grad_norm": 20.153647410714772,
+      "learning_rate": 4.634710743801653e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0901,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 0.06396484375,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -7.90625,
+      "step": 1043
+    },
+    {
+      "epoch": 1.2809815950920245,
+      "grad_norm": 6.924901551687326,
+      "learning_rate": 4.631404958677685e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -374.0,
+      "loss": 0.022,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -8.3125,
+      "step": 1044
+    },
+    {
+      "epoch": 1.2822085889570551,
+      "grad_norm": 12.95174867862595,
+      "learning_rate": 4.6280991735537184e-07,
+      "logits/chosen": -0.31640625,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0403,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.125,
+      "step": 1045
+    },
+    {
+      "epoch": 1.2834355828220858,
+      "grad_norm": 14.546745301507391,
+      "learning_rate": 4.6247933884297516e-07,
+      "logits/chosen": -0.099609375,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -374.0,
+      "loss": 0.063,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.83984375,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.53125,
+      "step": 1046
+    },
+    {
+      "epoch": 1.2846625766871167,
+      "grad_norm": 10.883403162683157,
+      "learning_rate": 4.621487603305785e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -376.0,
+      "loss": 0.035,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -8.625,
+      "step": 1047
+    },
+    {
+      "epoch": 1.2858895705521474,
+      "grad_norm": 8.026266419876157,
+      "learning_rate": 4.618181818181818e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -332.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0379,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.375,
+      "step": 1048
+    },
+    {
+      "epoch": 1.287116564417178,
+      "grad_norm": 9.940422173754929,
+      "learning_rate": 4.614876033057851e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0394,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.3125,
+      "step": 1049
+    },
+    {
+      "epoch": 1.2883435582822087,
+      "grad_norm": 13.289809124549075,
+      "learning_rate": 4.6115702479338843e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0543,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.25,
+      "step": 1050
+    },
+    {
+      "epoch": 1.2895705521472394,
+      "grad_norm": 8.246230191346283,
+      "learning_rate": 4.6082644628099175e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0325,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -8.875,
+      "step": 1051
+    },
+    {
+      "epoch": 1.29079754601227,
+      "grad_norm": 14.40820979773154,
+      "learning_rate": 4.6049586776859507e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0636,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -7.9375,
+      "step": 1052
+    },
+    {
+      "epoch": 1.2920245398773007,
+      "grad_norm": 11.868252173578338,
+      "learning_rate": 4.601652892561983e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0532,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.25,
+      "step": 1053
+    },
+    {
+      "epoch": 1.2932515337423314,
+      "grad_norm": 6.870935187145037,
+      "learning_rate": 4.598347107438016e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0361,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.4375,
+      "step": 1054
+    },
+    {
+      "epoch": 1.294478527607362,
+      "grad_norm": 10.48304399523796,
+      "learning_rate": 4.595041322314049e-07,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0388,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.8203125,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.875,
+      "step": 1055
+    },
+    {
+      "epoch": 1.2957055214723927,
+      "grad_norm": 4.362691885640252,
+      "learning_rate": 4.5917355371900823e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -322.0,
+      "loss": 0.0164,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.7734375,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.5,
+      "step": 1056
+    },
+    {
+      "epoch": 1.2969325153374234,
+      "grad_norm": 15.040919629551425,
+      "learning_rate": 4.5884297520661155e-07,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0419,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.375,
+      "step": 1057
+    },
+    {
+      "epoch": 1.298159509202454,
+      "grad_norm": 16.109841585162464,
+      "learning_rate": 4.5851239669421487e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0663,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -8.6875,
+      "step": 1058
+    },
+    {
+      "epoch": 1.2993865030674847,
+      "grad_norm": 7.994341213442368,
+      "learning_rate": 4.581818181818182e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0447,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.80859375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.875,
+      "step": 1059
+    },
+    {
+      "epoch": 1.3006134969325154,
+      "grad_norm": 14.221171845997816,
+      "learning_rate": 4.578512396694215e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -332.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0483,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.625,
+      "step": 1060
+    },
+    {
+      "epoch": 1.301840490797546,
+      "grad_norm": 11.819665312091933,
+      "learning_rate": 4.575206611570248e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0422,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.25,
+      "step": 1061
+    },
+    {
+      "epoch": 1.3030674846625767,
+      "grad_norm": 14.568973750937094,
+      "learning_rate": 4.5719008264462803e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -392.0,
+      "loss": 0.066,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.1875,
+      "step": 1062
+    },
+    {
+      "epoch": 1.3042944785276074,
+      "grad_norm": 9.167928996893854,
+      "learning_rate": 4.5685950413223135e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0537,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.5,
+      "step": 1063
+    },
+    {
+      "epoch": 1.305521472392638,
+      "grad_norm": 8.302137097278727,
+      "learning_rate": 4.5652892561983467e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0359,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.82421875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.0,
+      "step": 1064
+    },
+    {
+      "epoch": 1.3067484662576687,
+      "grad_norm": 10.525223064527983,
+      "learning_rate": 4.56198347107438e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -442.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0462,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7578125,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -9.0625,
+      "step": 1065
+    },
+    {
+      "epoch": 1.3079754601226994,
+      "grad_norm": 10.97929802196726,
+      "learning_rate": 4.558677685950413e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -456.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0383,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.3125,
+      "step": 1066
+    },
+    {
+      "epoch": 1.30920245398773,
+      "grad_norm": 10.890729674094034,
+      "learning_rate": 4.555371900826446e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -364.0,
+      "loss": 0.051,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.49609375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.0,
+      "step": 1067
+    },
+    {
+      "epoch": 1.3104294478527607,
+      "grad_norm": 13.051811444762572,
+      "learning_rate": 4.5520661157024794e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -340.0,
+      "loss": 0.062,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.5,
+      "step": 1068
+    },
+    {
+      "epoch": 1.3116564417177914,
+      "grad_norm": 11.077880050653722,
+      "learning_rate": 4.5487603305785126e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0566,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.625,
+      "step": 1069
+    },
+    {
+      "epoch": 1.312883435582822,
+      "grad_norm": 11.319123638531106,
+      "learning_rate": 4.5454545454545457e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0333,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.5625,
+      "step": 1070
+    },
+    {
+      "epoch": 1.3141104294478527,
+      "grad_norm": 15.58096616127429,
+      "learning_rate": 4.542148760330578e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0632,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.0,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.3125,
+      "step": 1071
+    },
+    {
+      "epoch": 1.3153374233128834,
+      "grad_norm": 9.810761244120958,
+      "learning_rate": 4.538842975206611e-07,
+      "logits/chosen": -0.1201171875,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0379,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -7.8125,
+      "step": 1072
+    },
+    {
+      "epoch": 1.316564417177914,
+      "grad_norm": 13.233835603974985,
+      "learning_rate": 4.535537190082644e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -324.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0544,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.90625,
+      "step": 1073
+    },
+    {
+      "epoch": 1.3177914110429447,
+      "grad_norm": 7.47915926353906,
+      "learning_rate": 4.5322314049586774e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0321,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.4375,
+      "step": 1074
+    },
+    {
+      "epoch": 1.3190184049079754,
+      "grad_norm": 11.99553982285267,
+      "learning_rate": 4.5289256198347106e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0402,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.125,
+      "step": 1075
+    },
+    {
+      "epoch": 1.320245398773006,
+      "grad_norm": 9.482049024318554,
+      "learning_rate": 4.5256198347107437e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0406,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.5,
+      "step": 1076
+    },
+    {
+      "epoch": 1.3214723926380367,
+      "grad_norm": 9.407916883456961,
+      "learning_rate": 4.522314049586777e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -294.0,
+      "logps/rejected": -290.0,
+      "loss": 0.0463,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.390625,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -8.125,
+      "step": 1077
+    },
+    {
+      "epoch": 1.3226993865030674,
+      "grad_norm": 12.437035820876217,
+      "learning_rate": 4.51900826446281e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.44921875,
+      "rewards/margins": 7.84375,
+      "rewards/rejected": -7.40625,
+      "step": 1078
+    },
+    {
+      "epoch": 1.323926380368098,
+      "grad_norm": 13.80505764294533,
+      "learning_rate": 4.515702479338843e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0704,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.4375,
+      "step": 1079
+    },
+    {
+      "epoch": 1.3251533742331287,
+      "grad_norm": 12.147288084424153,
+      "learning_rate": 4.5123966942148754e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0568,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.75,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -7.4375,
+      "step": 1080
+    },
+    {
+      "epoch": 1.3263803680981594,
+      "grad_norm": 7.917444989605004,
+      "learning_rate": 4.5090909090909086e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0379,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.89453125,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.1875,
+      "step": 1081
+    },
+    {
+      "epoch": 1.32760736196319,
+      "grad_norm": 8.384496689172522,
+      "learning_rate": 4.505785123966942e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0376,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.6875,
+      "step": 1082
+    },
+    {
+      "epoch": 1.3288343558282207,
+      "grad_norm": 9.161583572655749,
+      "learning_rate": 4.502479338842975e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -306.0,
+      "loss": 0.043,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.5390625,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -7.625,
+      "step": 1083
+    },
+    {
+      "epoch": 1.3300613496932514,
+      "grad_norm": 18.270345545412372,
+      "learning_rate": 4.499173553719008e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0751,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.78125,
+      "step": 1084
+    },
+    {
+      "epoch": 1.331288343558282,
+      "grad_norm": 13.799249791090745,
+      "learning_rate": 4.4958677685950413e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.404296875,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -7.8125,
+      "step": 1085
+    },
+    {
+      "epoch": 1.332515337423313,
+      "grad_norm": 8.946002080660248,
+      "learning_rate": 4.4925619834710744e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -310.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0496,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.25,
+      "step": 1086
+    },
+    {
+      "epoch": 1.3337423312883436,
+      "grad_norm": 10.788580914750392,
+      "learning_rate": 4.4892561983471076e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -318.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1943359375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.3125,
+      "step": 1087
+    },
+    {
+      "epoch": 1.3349693251533743,
+      "grad_norm": 14.32023199723315,
+      "learning_rate": 4.485950413223141e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0476,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.3125,
+      "step": 1088
+    },
+    {
+      "epoch": 1.336196319018405,
+      "grad_norm": 11.907439115072256,
+      "learning_rate": 4.482644628099173e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0615,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.0673828125,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -8.25,
+      "step": 1089
+    },
+    {
+      "epoch": 1.3374233128834356,
+      "grad_norm": 12.503556120707168,
+      "learning_rate": 4.479338842975206e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0519,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.197265625,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -8.0625,
+      "step": 1090
+    },
+    {
+      "epoch": 1.3386503067484663,
+      "grad_norm": 18.17820402883285,
+      "learning_rate": 4.4760330578512393e-07,
+      "logits/chosen": -0.0947265625,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0573,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -7.71875,
+      "step": 1091
+    },
+    {
+      "epoch": 1.339877300613497,
+      "grad_norm": 12.556890306296957,
+      "learning_rate": 4.4727272727272724e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.0625,
+      "step": 1092
+    },
+    {
+      "epoch": 1.3411042944785276,
+      "grad_norm": 10.170118448319533,
+      "learning_rate": 4.4694214876033056e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0468,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7265625,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.3125,
+      "step": 1093
+    },
+    {
+      "epoch": 1.3423312883435583,
+      "grad_norm": 14.331382863628885,
+      "learning_rate": 4.466115702479339e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -400.0,
+      "loss": 0.052,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.4375,
+      "step": 1094
+    },
+    {
+      "epoch": 1.343558282208589,
+      "grad_norm": 8.788938438317068,
+      "learning_rate": 4.462809917355372e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -440.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0296,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -9.375,
+      "step": 1095
+    },
+    {
+      "epoch": 1.3447852760736196,
+      "grad_norm": 22.147235958677946,
+      "learning_rate": 4.459504132231405e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0749,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.1875,
+      "step": 1096
+    },
+    {
+      "epoch": 1.3460122699386503,
+      "grad_norm": 8.228501370409944,
+      "learning_rate": 4.4561983471074383e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -426.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0324,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.375,
+      "step": 1097
+    },
+    {
+      "epoch": 1.347239263803681,
+      "grad_norm": 10.392801869852612,
+      "learning_rate": 4.4528925619834705e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.671875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.0,
+      "step": 1098
+    },
+    {
+      "epoch": 1.3484662576687116,
+      "grad_norm": 19.360520432132724,
+      "learning_rate": 4.4495867768595036e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0913,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.322265625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.1875,
+      "step": 1099
+    },
+    {
+      "epoch": 1.3496932515337423,
+      "grad_norm": 10.418818451409532,
+      "learning_rate": 4.446280991735537e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.490234375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0385,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -8.5625,
+      "step": 1100
+    },
+    {
+      "epoch": 1.350920245398773,
+      "grad_norm": 12.995574614197842,
+      "learning_rate": 4.44297520661157e-07,
+      "logits/chosen": -0.08203125,
+      "logits/rejected": -0.2392578125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0568,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.0,
+      "step": 1101
+    },
+    {
+      "epoch": 1.3521472392638036,
+      "grad_norm": 12.498300123184615,
+      "learning_rate": 4.439669421487603e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0541,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -9.875,
+      "step": 1102
+    },
+    {
+      "epoch": 1.3533742331288343,
+      "grad_norm": 15.391881436542993,
+      "learning_rate": 4.4363636363636363e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0679,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.796875,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.1875,
+      "step": 1103
+    },
+    {
+      "epoch": 1.354601226993865,
+      "grad_norm": 14.536968024679194,
+      "learning_rate": 4.4330578512396695e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0651,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.51953125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.0,
+      "step": 1104
+    },
+    {
+      "epoch": 1.3558282208588956,
+      "grad_norm": 6.1991402014729875,
+      "learning_rate": 4.4297520661157027e-07,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0267,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -9.6875,
+      "step": 1105
+    },
+    {
+      "epoch": 1.3570552147239263,
+      "grad_norm": 9.510072203259035,
+      "learning_rate": 4.426446280991735e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0466,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1357421875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.5,
+      "step": 1106
+    },
+    {
+      "epoch": 1.358282208588957,
+      "grad_norm": 11.345500016334285,
+      "learning_rate": 4.423140495867768e-07,
+      "logits/chosen": -0.0986328125,
+      "logits/rejected": -0.21875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -360.0,
+      "loss": 0.045,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.578125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.1875,
+      "step": 1107
+    },
+    {
+      "epoch": 1.3595092024539877,
+      "grad_norm": 8.67082878371813,
+      "learning_rate": 4.419834710743801e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -442.0,
+      "logps/rejected": -388.0,
+      "loss": 0.035,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.4375,
+      "step": 1108
+    },
+    {
+      "epoch": 1.3607361963190185,
+      "grad_norm": 11.14953002163676,
+      "learning_rate": 4.4165289256198343e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0551,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.6015625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.75,
+      "step": 1109
+    },
+    {
+      "epoch": 1.3619631901840492,
+      "grad_norm": 15.6793165244425,
+      "learning_rate": 4.4132231404958675e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0635,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.75,
+      "step": 1110
+    },
+    {
+      "epoch": 1.3631901840490799,
+      "grad_norm": 7.626396120756006,
+      "learning_rate": 4.4099173553719007e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -434.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0359,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.470703125,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.25,
+      "step": 1111
+    },
+    {
+      "epoch": 1.3644171779141105,
+      "grad_norm": 10.029174865101465,
+      "learning_rate": 4.406611570247934e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0503,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.46484375,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.5,
+      "step": 1112
+    },
+    {
+      "epoch": 1.3656441717791412,
+      "grad_norm": 10.857796592085158,
+      "learning_rate": 4.403305785123967e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0512,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.5,
+      "step": 1113
+    },
+    {
+      "epoch": 1.3668711656441719,
+      "grad_norm": 15.765671968754052,
+      "learning_rate": 4.4e-07,
+      "logits/chosen": -0.3515625,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -344.0,
+      "loss": 0.056,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.0272216796875,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -9.125,
+      "step": 1114
+    },
+    {
+      "epoch": 1.3680981595092025,
+      "grad_norm": 11.976634672452079,
+      "learning_rate": 4.3966942148760323e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0489,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.83984375,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.0625,
+      "step": 1115
+    },
+    {
+      "epoch": 1.3693251533742332,
+      "grad_norm": 12.12629302690259,
+      "learning_rate": 4.3933884297520655e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0589,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.5703125,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.1875,
+      "step": 1116
+    },
+    {
+      "epoch": 1.3705521472392639,
+      "grad_norm": 9.786096050833946,
+      "learning_rate": 4.3900826446280987e-07,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0454,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.9375,
+      "step": 1117
+    },
+    {
+      "epoch": 1.3717791411042946,
+      "grad_norm": 12.232079190104093,
+      "learning_rate": 4.386776859504132e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0636,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.3671875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.6875,
+      "step": 1118
+    },
+    {
+      "epoch": 1.3730061349693252,
+      "grad_norm": 9.928765564322998,
+      "learning_rate": 4.383471074380165e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0526,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.4140625,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -8.0,
+      "step": 1119
+    },
+    {
+      "epoch": 1.3742331288343559,
+      "grad_norm": 10.258302380832172,
+      "learning_rate": 4.380165289256198e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0325,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.625,
+      "step": 1120
+    },
+    {
+      "epoch": 1.3754601226993866,
+      "grad_norm": 20.979255833463608,
+      "learning_rate": 4.3768595041322314e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0915,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.125,
+      "step": 1121
+    },
+    {
+      "epoch": 1.3766871165644172,
+      "grad_norm": 11.325879983054083,
+      "learning_rate": 4.3735537190082646e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0423,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2890625,
+      "rewards/margins": 8.3125,
+      "rewards/rejected": -8.0625,
+      "step": 1122
+    },
+    {
+      "epoch": 1.377914110429448,
+      "grad_norm": 10.020287970518874,
+      "learning_rate": 4.370247933884298e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -436.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0387,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.3515625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.1875,
+      "step": 1123
+    },
+    {
+      "epoch": 1.3791411042944786,
+      "grad_norm": 11.809777379798582,
+      "learning_rate": 4.36694214876033e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0466,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.56640625,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.4375,
+      "step": 1124
+    },
+    {
+      "epoch": 1.3803680981595092,
+      "grad_norm": 8.666119379064078,
+      "learning_rate": 4.363636363636363e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0368,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.484375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.1875,
+      "step": 1125
+    },
+    {
+      "epoch": 1.38159509202454,
+      "grad_norm": 9.38301821790362,
+      "learning_rate": 4.360330578512396e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0395,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.51953125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.375,
+      "step": 1126
+    },
+    {
+      "epoch": 1.3828220858895706,
+      "grad_norm": 13.160048641339023,
+      "learning_rate": 4.3570247933884294e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -442.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0599,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.90625,
+      "step": 1127
+    },
+    {
+      "epoch": 1.3840490797546012,
+      "grad_norm": 14.649656027167724,
+      "learning_rate": 4.3537190082644626e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0755,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.3203125,
+      "rewards/margins": 7.8125,
+      "rewards/rejected": -7.5,
+      "step": 1128
+    },
+    {
+      "epoch": 1.385276073619632,
+      "grad_norm": 8.607132085557556,
+      "learning_rate": 4.350413223140496e-07,
+      "logits/chosen": -0.11376953125,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0342,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.59375,
+      "step": 1129
+    },
+    {
+      "epoch": 1.3865030674846626,
+      "grad_norm": 18.504640307311753,
+      "learning_rate": 4.347107438016529e-07,
+      "logits/chosen": -0.0634765625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0563,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -7.75,
+      "step": 1130
+    },
+    {
+      "epoch": 1.3877300613496932,
+      "grad_norm": 9.275241570046726,
+      "learning_rate": 4.343801652892562e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -7.75,
+      "step": 1131
+    },
+    {
+      "epoch": 1.388957055214724,
+      "grad_norm": 13.545233665904872,
+      "learning_rate": 4.3404958677685953e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -484.0,
+      "logps/rejected": -380.0,
+      "loss": 0.042,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -7.46875,
+      "step": 1132
+    },
+    {
+      "epoch": 1.3901840490797546,
+      "grad_norm": 9.02011133210679,
+      "learning_rate": 4.3371900826446274e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0402,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.0,
+      "step": 1133
+    },
+    {
+      "epoch": 1.3914110429447852,
+      "grad_norm": 9.626052329267141,
+      "learning_rate": 4.3338842975206606e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -316.0,
+      "logps/rejected": -312.0,
+      "loss": 0.0531,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.0732421875,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -8.1875,
+      "step": 1134
+    },
+    {
+      "epoch": 1.392638036809816,
+      "grad_norm": 9.379296233151626,
+      "learning_rate": 4.330578512396694e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.0,
+      "step": 1135
+    },
+    {
+      "epoch": 1.3938650306748466,
+      "grad_norm": 18.617309825421266,
+      "learning_rate": 4.327272727272727e-07,
+      "logits/chosen": -0.357421875,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -458.0,
+      "logps/rejected": -380.0,
+      "loss": 0.067,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.9375,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -8.1875,
+      "step": 1136
+    },
+    {
+      "epoch": 1.3950920245398772,
+      "grad_norm": 16.539416510452526,
+      "learning_rate": 4.32396694214876e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.55078125,
+      "logps/chosen": -458.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0659,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.6875,
+      "step": 1137
+    },
+    {
+      "epoch": 1.396319018404908,
+      "grad_norm": 9.959030275396955,
+      "learning_rate": 4.3206611570247933e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0397,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.0625,
+      "step": 1138
+    },
+    {
+      "epoch": 1.3975460122699386,
+      "grad_norm": 10.991054262003571,
+      "learning_rate": 4.3173553719008265e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.490234375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0449,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -7.9375,
+      "step": 1139
+    },
+    {
+      "epoch": 1.3987730061349692,
+      "grad_norm": 13.764692227891263,
+      "learning_rate": 4.3140495867768597e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -342.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0488,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.72265625,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.96875,
+      "step": 1140
+    },
+    {
+      "epoch": 1.4,
+      "grad_norm": 9.018367038711595,
+      "learning_rate": 4.310743801652893e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -332.0,
+      "loss": 0.032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.828125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.84375,
+      "step": 1141
+    },
+    {
+      "epoch": 1.4012269938650306,
+      "grad_norm": 14.08834337812807,
+      "learning_rate": 4.307438016528925e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0521,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.875,
+      "step": 1142
+    },
+    {
+      "epoch": 1.4024539877300612,
+      "grad_norm": 10.911725420548523,
+      "learning_rate": 4.304132231404958e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0453,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.515625,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.3125,
+      "step": 1143
+    },
+    {
+      "epoch": 1.403680981595092,
+      "grad_norm": 10.875646708704325,
+      "learning_rate": 4.3008264462809913e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0366,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.89453125,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.1875,
+      "step": 1144
+    },
+    {
+      "epoch": 1.4049079754601226,
+      "grad_norm": 19.771180533625117,
+      "learning_rate": 4.2975206611570245e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -452.0,
+      "logps/rejected": -402.0,
+      "loss": 0.055,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.84375,
+      "step": 1145
+    },
+    {
+      "epoch": 1.4061349693251532,
+      "grad_norm": 11.6301527103396,
+      "learning_rate": 4.2942148760330577e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0461,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.4375,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.625,
+      "step": 1146
+    },
+    {
+      "epoch": 1.407361963190184,
+      "grad_norm": 4.878005421347195,
+      "learning_rate": 4.290909090909091e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0252,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.34375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.6875,
+      "step": 1147
+    },
+    {
+      "epoch": 1.4085889570552146,
+      "grad_norm": 8.16557883584569,
+      "learning_rate": 4.287603305785124e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0307,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.5,
+      "step": 1148
+    },
+    {
+      "epoch": 1.4098159509202455,
+      "grad_norm": 8.164045264121068,
+      "learning_rate": 4.284297520661157e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0278,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.9375,
+      "step": 1149
+    },
+    {
+      "epoch": 1.4110429447852761,
+      "grad_norm": 10.04611876987732,
+      "learning_rate": 4.2809917355371904e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0408,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.462890625,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.625,
+      "step": 1150
+    },
+    {
+      "epoch": 1.4122699386503068,
+      "grad_norm": 10.11694907782239,
+      "learning_rate": 4.2776859504132225e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0399,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.482421875,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.875,
+      "step": 1151
+    },
+    {
+      "epoch": 1.4134969325153375,
+      "grad_norm": 6.802420223259662,
+      "learning_rate": 4.2743801652892557e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0252,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.0,
+      "step": 1152
+    },
+    {
+      "epoch": 1.4147239263803681,
+      "grad_norm": 12.518162366564754,
+      "learning_rate": 4.271074380165289e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0518,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -8.8125,
+      "step": 1153
+    },
+    {
+      "epoch": 1.4159509202453988,
+      "grad_norm": 14.27379900942779,
+      "learning_rate": 4.267768595041322e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.470703125,
+      "logps/chosen": -442.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0425,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.1875,
+      "step": 1154
+    },
+    {
+      "epoch": 1.4171779141104295,
+      "grad_norm": 10.309850585087835,
+      "learning_rate": 4.264462809917355e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0442,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1865234375,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.8125,
+      "step": 1155
+    },
+    {
+      "epoch": 1.4184049079754601,
+      "grad_norm": 15.590234943117638,
+      "learning_rate": 4.2611570247933884e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -434.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0428,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -9.5,
+      "step": 1156
+    },
+    {
+      "epoch": 1.4196319018404908,
+      "grad_norm": 9.987550517532815,
+      "learning_rate": 4.2578512396694215e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -446.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0317,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -8.9375,
+      "step": 1157
+    },
+    {
+      "epoch": 1.4208588957055215,
+      "grad_norm": 7.931416115322814,
+      "learning_rate": 4.2545454545454547e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -342.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0371,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.828125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.0625,
+      "step": 1158
+    },
+    {
+      "epoch": 1.4220858895705522,
+      "grad_norm": 18.05171696034389,
+      "learning_rate": 4.251239669421488e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0767,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.296875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.8125,
+      "step": 1159
+    },
+    {
+      "epoch": 1.4233128834355828,
+      "grad_norm": 13.943376246751296,
+      "learning_rate": 4.24793388429752e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0449,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.65625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.875,
+      "step": 1160
+    },
+    {
+      "epoch": 1.4245398773006135,
+      "grad_norm": 12.317132054833733,
+      "learning_rate": 4.244628099173553e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0585,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.75,
+      "step": 1161
+    },
+    {
+      "epoch": 1.4257668711656442,
+      "grad_norm": 9.354899646749724,
+      "learning_rate": 4.2413223140495864e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0414,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.0,
+      "step": 1162
+    },
+    {
+      "epoch": 1.4269938650306748,
+      "grad_norm": 8.354103036645931,
+      "learning_rate": 4.2380165289256195e-07,
+      "logits/chosen": -0.353515625,
+      "logits/rejected": -0.59765625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0303,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.4375,
+      "step": 1163
+    },
+    {
+      "epoch": 1.4282208588957055,
+      "grad_norm": 17.92923732721561,
+      "learning_rate": 4.2347107438016527e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.24609375,
+      "logps/chosen": -330.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0599,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.53125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.75,
+      "step": 1164
+    },
+    {
+      "epoch": 1.4294478527607362,
+      "grad_norm": 8.92045181790276,
+      "learning_rate": 4.231404958677686e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0279,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.3671875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.125,
+      "step": 1165
+    },
+    {
+      "epoch": 1.4306748466257668,
+      "grad_norm": 9.823343955782265,
+      "learning_rate": 4.228099173553719e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0317,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.5625,
+      "step": 1166
+    },
+    {
+      "epoch": 1.4319018404907975,
+      "grad_norm": 8.303351669646478,
+      "learning_rate": 4.224793388429752e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0256,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.6328125,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.6875,
+      "step": 1167
+    },
+    {
+      "epoch": 1.4331288343558282,
+      "grad_norm": 13.096682943463716,
+      "learning_rate": 4.2214876033057854e-07,
+      "logits/chosen": -0.248046875,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0432,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.3359375,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.8125,
+      "step": 1168
+    },
+    {
+      "epoch": 1.4343558282208588,
+      "grad_norm": 8.383755609382192,
+      "learning_rate": 4.2181818181818175e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0339,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.71875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.25,
+      "step": 1169
+    },
+    {
+      "epoch": 1.4355828220858895,
+      "grad_norm": 13.087573566535397,
+      "learning_rate": 4.2148760330578507e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0444,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.443359375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.1875,
+      "step": 1170
+    },
+    {
+      "epoch": 1.4368098159509202,
+      "grad_norm": 11.843632840671855,
+      "learning_rate": 4.211570247933884e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0424,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -8.9375,
+      "step": 1171
+    },
+    {
+      "epoch": 1.438036809815951,
+      "grad_norm": 13.430060643840084,
+      "learning_rate": 4.208264462809917e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0613,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.287109375,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -8.0625,
+      "step": 1172
+    },
+    {
+      "epoch": 1.4392638036809817,
+      "grad_norm": 13.140000801988304,
+      "learning_rate": 4.20495867768595e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0435,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.61328125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.9375,
+      "step": 1173
+    },
+    {
+      "epoch": 1.4404907975460124,
+      "grad_norm": 10.037701828497555,
+      "learning_rate": 4.2016528925619834e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0387,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.3125,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.75,
+      "step": 1174
+    },
+    {
+      "epoch": 1.441717791411043,
+      "grad_norm": 7.925062400196428,
+      "learning_rate": 4.1983471074380166e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -342.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0409,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.12890625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.75,
+      "step": 1175
+    },
+    {
+      "epoch": 1.4429447852760737,
+      "grad_norm": 7.918415538616964,
+      "learning_rate": 4.19504132231405e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0375,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1015625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.5,
+      "step": 1176
+    },
+    {
+      "epoch": 1.4441717791411044,
+      "grad_norm": 9.953622472970856,
+      "learning_rate": 4.191735537190083e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0429,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.0,
+      "step": 1177
+    },
+    {
+      "epoch": 1.445398773006135,
+      "grad_norm": 9.134426899002403,
+      "learning_rate": 4.188429752066115e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0339,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -8.5,
+      "step": 1178
+    },
+    {
+      "epoch": 1.4466257668711657,
+      "grad_norm": 9.094305897866864,
+      "learning_rate": 4.185123966942148e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0398,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.625,
+      "step": 1179
+    },
+    {
+      "epoch": 1.4478527607361964,
+      "grad_norm": 13.578461286413251,
+      "learning_rate": 4.1818181818181814e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0506,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.46484375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.875,
+      "step": 1180
+    },
+    {
+      "epoch": 1.449079754601227,
+      "grad_norm": 7.992882025383505,
+      "learning_rate": 4.1785123966942146e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0253,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.9375,
+      "step": 1181
+    },
+    {
+      "epoch": 1.4503067484662577,
+      "grad_norm": 10.0106751323216,
+      "learning_rate": 4.175206611570248e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -384.0,
+      "loss": 0.036,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.21875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.625,
+      "step": 1182
+    },
+    {
+      "epoch": 1.4515337423312884,
+      "grad_norm": 15.000115439366018,
+      "learning_rate": 4.171900826446281e-07,
+      "logits/chosen": -0.24609375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0654,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.193359375,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -8.0,
+      "step": 1183
+    },
+    {
+      "epoch": 1.452760736196319,
+      "grad_norm": 10.305727034916446,
+      "learning_rate": 4.168595041322314e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0461,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.72265625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.8125,
+      "step": 1184
+    },
+    {
+      "epoch": 1.4539877300613497,
+      "grad_norm": 10.606454430153265,
+      "learning_rate": 4.1652892561983473e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0407,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.3125,
+      "step": 1185
+    },
+    {
+      "epoch": 1.4552147239263804,
+      "grad_norm": 9.636022481666009,
+      "learning_rate": 4.1619834710743805e-07,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0559,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.96875,
+      "step": 1186
+    },
+    {
+      "epoch": 1.456441717791411,
+      "grad_norm": 11.476673812202115,
+      "learning_rate": 4.1586776859504126e-07,
+      "logits/chosen": -0.2421875,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -450.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0466,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.625,
+      "step": 1187
+    },
+    {
+      "epoch": 1.4576687116564417,
+      "grad_norm": 11.105210699415723,
+      "learning_rate": 4.155371900826446e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.9375,
+      "step": 1188
+    },
+    {
+      "epoch": 1.4588957055214724,
+      "grad_norm": 12.661874309989074,
+      "learning_rate": 4.152066115702479e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0575,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.125,
+      "step": 1189
+    },
+    {
+      "epoch": 1.460122699386503,
+      "grad_norm": 8.654978193473417,
+      "learning_rate": 4.148760330578512e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0357,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.5625,
+      "step": 1190
+    },
+    {
+      "epoch": 1.4613496932515337,
+      "grad_norm": 11.584776568508845,
+      "learning_rate": 4.1454545454545453e-07,
+      "logits/chosen": -0.11376953125,
+      "logits/rejected": -0.25,
+      "logps/chosen": -364.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0451,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.75,
+      "step": 1191
+    },
+    {
+      "epoch": 1.4625766871165644,
+      "grad_norm": 11.938546188570182,
+      "learning_rate": 4.1421487603305785e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0479,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.75,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.375,
+      "step": 1192
+    },
+    {
+      "epoch": 1.463803680981595,
+      "grad_norm": 8.57539996640767,
+      "learning_rate": 4.1388429752066117e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -452.0,
+      "logps/rejected": -390.0,
+      "loss": 0.029,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.0625,
+      "step": 1193
+    },
+    {
+      "epoch": 1.4650306748466257,
+      "grad_norm": 9.189664993288334,
+      "learning_rate": 4.135537190082645e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -382.0,
+      "loss": 0.021,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.375,
+      "step": 1194
+    },
+    {
+      "epoch": 1.4662576687116564,
+      "grad_norm": 9.081781667653363,
+      "learning_rate": 4.132231404958677e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -384.0,
+      "loss": 0.04,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.76171875,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.25,
+      "step": 1195
+    },
+    {
+      "epoch": 1.467484662576687,
+      "grad_norm": 12.415433200199677,
+      "learning_rate": 4.12892561983471e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0524,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.4453125,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.96875,
+      "step": 1196
+    },
+    {
+      "epoch": 1.4687116564417177,
+      "grad_norm": 9.943422765866982,
+      "learning_rate": 4.1256198347107433e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0466,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.25,
+      "step": 1197
+    },
+    {
+      "epoch": 1.4699386503067484,
+      "grad_norm": 14.559683333211177,
+      "learning_rate": 4.1223140495867765e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0527,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.73828125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.96875,
+      "step": 1198
+    },
+    {
+      "epoch": 1.471165644171779,
+      "grad_norm": 17.776993736511606,
+      "learning_rate": 4.1190082644628097e-07,
+      "logits/chosen": -0.31640625,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0515,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.6875,
+      "step": 1199
+    },
+    {
+      "epoch": 1.4723926380368098,
+      "grad_norm": 7.694268373865595,
+      "learning_rate": 4.115702479338843e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0276,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.55859375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.9375,
+      "step": 1200
+    },
+    {
+      "epoch": 1.4736196319018404,
+      "grad_norm": 7.758540238755445,
+      "learning_rate": 4.112396694214876e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0271,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.25,
+      "step": 1201
+    },
+    {
+      "epoch": 1.474846625766871,
+      "grad_norm": 9.980200405050743,
+      "learning_rate": 4.109090909090909e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0357,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.375,
+      "step": 1202
+    },
+    {
+      "epoch": 1.4760736196319018,
+      "grad_norm": 8.852224972647532,
+      "learning_rate": 4.1057851239669424e-07,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0434,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.66015625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.4375,
+      "step": 1203
+    },
+    {
+      "epoch": 1.4773006134969324,
+      "grad_norm": 10.918645848118,
+      "learning_rate": 4.1024793388429745e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0408,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.4375,
+      "step": 1204
+    },
+    {
+      "epoch": 1.478527607361963,
+      "grad_norm": 15.360319822899514,
+      "learning_rate": 4.0991735537190077e-07,
+      "logits/chosen": -0.1064453125,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0614,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.1875,
+      "step": 1205
+    },
+    {
+      "epoch": 1.4797546012269938,
+      "grad_norm": 6.938456095237597,
+      "learning_rate": 4.095867768595041e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -344.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0355,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.55078125,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.3125,
+      "step": 1206
+    },
+    {
+      "epoch": 1.4809815950920244,
+      "grad_norm": 12.838247596989346,
+      "learning_rate": 4.092561983471074e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0441,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.296875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.5,
+      "step": 1207
+    },
+    {
+      "epoch": 1.482208588957055,
+      "grad_norm": 14.774384990299971,
+      "learning_rate": 4.089256198347107e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0675,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.294921875,
+      "rewards/margins": 8.3125,
+      "rewards/rejected": -8.0625,
+      "step": 1208
+    },
+    {
+      "epoch": 1.4834355828220858,
+      "grad_norm": 11.62853808692027,
+      "learning_rate": 4.0859504132231404e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0536,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.47265625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.875,
+      "step": 1209
+    },
+    {
+      "epoch": 1.4846625766871164,
+      "grad_norm": 11.857036692599188,
+      "learning_rate": 4.0826446280991736e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0305,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.85546875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.5625,
+      "step": 1210
+    },
+    {
+      "epoch": 1.4858895705521473,
+      "grad_norm": 4.69235230102138,
+      "learning_rate": 4.079338842975207e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -360.0,
+      "loss": 0.019,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.625,
+      "step": 1211
+    },
+    {
+      "epoch": 1.487116564417178,
+      "grad_norm": 9.185641508263624,
+      "learning_rate": 4.07603305785124e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0428,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.326171875,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.9375,
+      "step": 1212
+    },
+    {
+      "epoch": 1.4883435582822087,
+      "grad_norm": 8.195477892364691,
+      "learning_rate": 4.072727272727272e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.55078125,
+      "logps/chosen": -428.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0365,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.75,
+      "step": 1213
+    },
+    {
+      "epoch": 1.4895705521472393,
+      "grad_norm": 14.570501254636081,
+      "learning_rate": 4.069421487603305e-07,
+      "logits/chosen": -0.2421875,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -346.0,
+      "logps/rejected": -356.0,
+      "loss": 0.056,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": -0.018310546875,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -9.0625,
+      "step": 1214
+    },
+    {
+      "epoch": 1.49079754601227,
+      "grad_norm": 11.800836028244067,
+      "learning_rate": 4.0661157024793384e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -378.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0324,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.5625,
+      "step": 1215
+    },
+    {
+      "epoch": 1.4920245398773007,
+      "grad_norm": 12.335957050020774,
+      "learning_rate": 4.0628099173553716e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0482,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.00836181640625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.625,
+      "step": 1216
+    },
+    {
+      "epoch": 1.4932515337423313,
+      "grad_norm": 9.006796955903845,
+      "learning_rate": 4.059504132231405e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -382.0,
+      "loss": 0.03,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.6875,
+      "step": 1217
+    },
+    {
+      "epoch": 1.494478527607362,
+      "grad_norm": 8.123325606423187,
+      "learning_rate": 4.056198347107438e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0321,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.01385498046875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.9375,
+      "step": 1218
+    },
+    {
+      "epoch": 1.4957055214723927,
+      "grad_norm": 12.558099415012947,
+      "learning_rate": 4.052892561983471e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -388.0,
+      "loss": 0.036,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.4296875,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.8125,
+      "step": 1219
+    },
+    {
+      "epoch": 1.4969325153374233,
+      "grad_norm": 12.775377284659152,
+      "learning_rate": 4.0495867768595043e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0506,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.2734375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.875,
+      "step": 1220
+    },
+    {
+      "epoch": 1.498159509202454,
+      "grad_norm": 11.306165820615583,
+      "learning_rate": 4.0462809917355375e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0388,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.71875,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.125,
+      "step": 1221
+    },
+    {
+      "epoch": 1.4993865030674847,
+      "grad_norm": 12.268638749924271,
+      "learning_rate": 4.0429752066115696e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -334.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.33984375,
+      "rewards/margins": 8.3125,
+      "rewards/rejected": -8.625,
+      "step": 1222
+    },
+    {
+      "epoch": 1.5006134969325153,
+      "grad_norm": 16.447342830994092,
+      "learning_rate": 4.039669421487603e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -378.0,
+      "loss": 0.049,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.5625,
+      "step": 1223
+    },
+    {
+      "epoch": 1.501840490797546,
+      "grad_norm": 10.490780921280326,
+      "learning_rate": 4.036363636363636e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0536,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.458984375,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -8.4375,
+      "step": 1224
+    },
+    {
+      "epoch": 1.5030674846625767,
+      "grad_norm": 8.180288581649103,
+      "learning_rate": 4.033057851239669e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0337,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.33203125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.5625,
+      "step": 1225
+    },
+    {
+      "epoch": 1.5042944785276073,
+      "grad_norm": 11.757770803193841,
+      "learning_rate": 4.0297520661157023e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -358.0,
+      "loss": 0.035,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.0179443359375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.6875,
+      "step": 1226
+    },
+    {
+      "epoch": 1.505521472392638,
+      "grad_norm": 7.402348890852564,
+      "learning_rate": 4.0264462809917355e-07,
+      "logits/chosen": -0.38671875,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0259,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.625,
+      "step": 1227
+    },
+    {
+      "epoch": 1.5067484662576687,
+      "grad_norm": 24.868807973648707,
+      "learning_rate": 4.0231404958677686e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -400.0,
+      "loss": 0.075,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.486328125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.25,
+      "step": 1228
+    },
+    {
+      "epoch": 1.5079754601226993,
+      "grad_norm": 10.416352098408494,
+      "learning_rate": 4.019834710743802e-07,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0321,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.330078125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.6875,
+      "step": 1229
+    },
+    {
+      "epoch": 1.50920245398773,
+      "grad_norm": 12.727906851382768,
+      "learning_rate": 4.016528925619835e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0488,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.74609375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.96875,
+      "step": 1230
+    },
+    {
+      "epoch": 1.510429447852761,
+      "grad_norm": 7.3509385262124125,
+      "learning_rate": 4.013223140495867e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -348.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0284,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.75390625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.78125,
+      "step": 1231
+    },
+    {
+      "epoch": 1.5116564417177916,
+      "grad_norm": 9.498443490772354,
+      "learning_rate": 4.0099173553719003e-07,
+      "logits/chosen": -0.37109375,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0265,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.2109375,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -8.8125,
+      "step": 1232
+    },
+    {
+      "epoch": 1.5128834355828222,
+      "grad_norm": 9.88336081694715,
+      "learning_rate": 4.0066115702479335e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0359,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.375,
+      "step": 1233
+    },
+    {
+      "epoch": 1.514110429447853,
+      "grad_norm": 20.777045013839473,
+      "learning_rate": 4.0033057851239666e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0549,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.84375,
+      "step": 1234
+    },
+    {
+      "epoch": 1.5153374233128836,
+      "grad_norm": 11.047074969488666,
+      "learning_rate": 4e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0423,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.171875,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -8.0,
+      "step": 1235
+    },
+    {
+      "epoch": 1.5165644171779142,
+      "grad_norm": 13.010857169907007,
+      "learning_rate": 3.996694214876033e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0398,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.71875,
+      "step": 1236
+    },
+    {
+      "epoch": 1.517791411042945,
+      "grad_norm": 14.316784677028014,
+      "learning_rate": 3.993388429752066e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.5,
+      "logps/chosen": -424.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0461,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.90234375,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.0625,
+      "step": 1237
+    },
+    {
+      "epoch": 1.5190184049079756,
+      "grad_norm": 10.925753336150258,
+      "learning_rate": 3.990082644628099e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -458.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0332,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.125,
+      "step": 1238
+    },
+    {
+      "epoch": 1.5202453987730062,
+      "grad_norm": 14.155768697592602,
+      "learning_rate": 3.986776859504132e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0563,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.9375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.625,
+      "step": 1239
+    },
+    {
+      "epoch": 1.521472392638037,
+      "grad_norm": 12.010081911987827,
+      "learning_rate": 3.983471074380165e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0542,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.68359375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.3125,
+      "step": 1240
+    },
+    {
+      "epoch": 1.5226993865030676,
+      "grad_norm": 15.161747078373114,
+      "learning_rate": 3.9801652892561984e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0635,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.96875,
+      "step": 1241
+    },
+    {
+      "epoch": 1.5239263803680982,
+      "grad_norm": 18.12655990544915,
+      "learning_rate": 3.976859504132231e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0519,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.359375,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.25,
+      "step": 1242
+    },
+    {
+      "epoch": 1.525153374233129,
+      "grad_norm": 12.22688926301596,
+      "learning_rate": 3.973553719008264e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -326.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0418,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.11376953125,
+      "rewards/margins": 8.3125,
+      "rewards/rejected": -8.1875,
+      "step": 1243
+    },
+    {
+      "epoch": 1.5263803680981596,
+      "grad_norm": 7.942928883549814,
+      "learning_rate": 3.9702479338842974e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -336.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0312,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1552734375,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -8.0625,
+      "step": 1244
+    },
+    {
+      "epoch": 1.5276073619631902,
+      "grad_norm": 15.392015274267582,
+      "learning_rate": 3.9669421487603305e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0783,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.6796875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.9375,
+      "step": 1245
+    },
+    {
+      "epoch": 1.528834355828221,
+      "grad_norm": 6.4157878266794315,
+      "learning_rate": 3.9636363636363637e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0324,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.482421875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.25,
+      "step": 1246
+    },
+    {
+      "epoch": 1.5300613496932516,
+      "grad_norm": 6.140011997029994,
+      "learning_rate": 3.9603305785123964e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0233,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.85546875,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.4375,
+      "step": 1247
+    },
+    {
+      "epoch": 1.5312883435582823,
+      "grad_norm": 8.639626779722814,
+      "learning_rate": 3.9570247933884295e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0254,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.8828125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.0,
+      "step": 1248
+    },
+    {
+      "epoch": 1.532515337423313,
+      "grad_norm": 15.793514748193386,
+      "learning_rate": 3.9537190082644627e-07,
+      "logits/chosen": -0.06884765625,
+      "logits/rejected": -0.1806640625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0715,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.09912109375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -9.125,
+      "step": 1249
+    },
+    {
+      "epoch": 1.5337423312883436,
+      "grad_norm": 12.584742135566902,
+      "learning_rate": 3.950413223140496e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -442.0,
+      "logps/rejected": -446.0,
+      "loss": 0.0431,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.81640625,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -10.0625,
+      "step": 1250
+    },
+    {
+      "epoch": 1.5349693251533743,
+      "grad_norm": 16.598890486502036,
+      "learning_rate": 3.9471074380165285e-07,
+      "logits/chosen": -0.37109375,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0551,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -9.6875,
+      "step": 1251
+    },
+    {
+      "epoch": 1.536196319018405,
+      "grad_norm": 14.5035620684537,
+      "learning_rate": 3.9438016528925617e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -352.0,
+      "loss": 0.053,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.3828125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -10.0625,
+      "step": 1252
+    },
+    {
+      "epoch": 1.5374233128834356,
+      "grad_norm": 13.289899357850107,
+      "learning_rate": 3.940495867768595e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0467,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.76953125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.125,
+      "step": 1253
+    },
+    {
+      "epoch": 1.5386503067484663,
+      "grad_norm": 13.876288513389907,
+      "learning_rate": 3.937190082644628e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0566,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.9375,
+      "step": 1254
+    },
+    {
+      "epoch": 1.539877300613497,
+      "grad_norm": 10.631076981710617,
+      "learning_rate": 3.933884297520661e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0369,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.859375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.0625,
+      "step": 1255
+    },
+    {
+      "epoch": 1.5411042944785276,
+      "grad_norm": 13.82590794396477,
+      "learning_rate": 3.930578512396694e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0441,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.5,
+      "step": 1256
+    },
+    {
+      "epoch": 1.5423312883435583,
+      "grad_norm": 14.32921649123448,
+      "learning_rate": 3.927272727272727e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0511,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.400390625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.875,
+      "step": 1257
+    },
+    {
+      "epoch": 1.543558282208589,
+      "grad_norm": 10.584191334594715,
+      "learning_rate": 3.92396694214876e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0399,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.625,
+      "step": 1258
+    },
+    {
+      "epoch": 1.5447852760736196,
+      "grad_norm": 10.748301553600525,
+      "learning_rate": 3.9206611570247934e-07,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0365,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.78515625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.8125,
+      "step": 1259
+    },
+    {
+      "epoch": 1.5460122699386503,
+      "grad_norm": 17.235825888413157,
+      "learning_rate": 3.917355371900826e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -450.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0427,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.9765625,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.9375,
+      "step": 1260
+    },
+    {
+      "epoch": 1.547239263803681,
+      "grad_norm": 10.355163477842352,
+      "learning_rate": 3.914049586776859e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0339,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.56640625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.3125,
+      "step": 1261
+    },
+    {
+      "epoch": 1.5484662576687116,
+      "grad_norm": 11.04210181900335,
+      "learning_rate": 3.9107438016528924e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0509,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.267578125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.6875,
+      "step": 1262
+    },
+    {
+      "epoch": 1.5496932515337423,
+      "grad_norm": 10.926373739741395,
+      "learning_rate": 3.9074380165289256e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0472,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.5625,
+      "step": 1263
+    },
+    {
+      "epoch": 1.550920245398773,
+      "grad_norm": 15.760559572323228,
+      "learning_rate": 3.904132231404958e-07,
+      "logits/chosen": -0.11474609375,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0745,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.453125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.4375,
+      "step": 1264
+    },
+    {
+      "epoch": 1.5521472392638036,
+      "grad_norm": 8.333997789167297,
+      "learning_rate": 3.9008264462809914e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0378,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.51171875,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.75,
+      "step": 1265
+    },
+    {
+      "epoch": 1.5533742331288343,
+      "grad_norm": 11.756550407636311,
+      "learning_rate": 3.8975206611570246e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0417,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.255859375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.4375,
+      "step": 1266
+    },
+    {
+      "epoch": 1.554601226993865,
+      "grad_norm": 16.803054241328326,
+      "learning_rate": 3.894214876033058e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -344.0,
+      "loss": 0.095,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 0.80859375,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.375,
+      "step": 1267
+    },
+    {
+      "epoch": 1.5558282208588956,
+      "grad_norm": 9.056787760649906,
+      "learning_rate": 3.890909090909091e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -496.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0293,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2109375,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.5625,
+      "step": 1268
+    },
+    {
+      "epoch": 1.5570552147239263,
+      "grad_norm": 11.529113923647646,
+      "learning_rate": 3.8876033057851236e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0361,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.4375,
+      "step": 1269
+    },
+    {
+      "epoch": 1.558282208588957,
+      "grad_norm": 12.885170546155125,
+      "learning_rate": 3.884297520661157e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0353,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.67578125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.1875,
+      "step": 1270
+    },
+    {
+      "epoch": 1.5595092024539876,
+      "grad_norm": 9.512525534029331,
+      "learning_rate": 3.88099173553719e-07,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.26953125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0345,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.65625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.25,
+      "step": 1271
+    },
+    {
+      "epoch": 1.5607361963190183,
+      "grad_norm": 5.776653205274608,
+      "learning_rate": 3.877685950413223e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0195,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.3125,
+      "step": 1272
+    },
+    {
+      "epoch": 1.561963190184049,
+      "grad_norm": 18.75333767957364,
+      "learning_rate": 3.874380165289256e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0867,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.625,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.125,
+      "step": 1273
+    },
+    {
+      "epoch": 1.5631901840490796,
+      "grad_norm": 6.671859500637093,
+      "learning_rate": 3.871074380165289e-07,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0226,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.4375,
+      "step": 1274
+    },
+    {
+      "epoch": 1.5644171779141103,
+      "grad_norm": 14.51239466060054,
+      "learning_rate": 3.867768595041322e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0681,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.419921875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.4375,
+      "step": 1275
+    },
+    {
+      "epoch": 1.565644171779141,
+      "grad_norm": 6.983571368573247,
+      "learning_rate": 3.8644628099173553e-07,
+      "logits/chosen": -0.12109375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0252,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.1875,
+      "step": 1276
+    },
+    {
+      "epoch": 1.5668711656441716,
+      "grad_norm": 20.362210341411654,
+      "learning_rate": 3.8611570247933885e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0691,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.431640625,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -7.84375,
+      "step": 1277
+    },
+    {
+      "epoch": 1.5680981595092025,
+      "grad_norm": 9.297990677048228,
+      "learning_rate": 3.857851239669421e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0298,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.4296875,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -8.75,
+      "step": 1278
+    },
+    {
+      "epoch": 1.5693251533742332,
+      "grad_norm": 13.341995153514718,
+      "learning_rate": 3.8545454545454543e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0559,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.6328125,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.78125,
+      "step": 1279
+    },
+    {
+      "epoch": 1.5705521472392638,
+      "grad_norm": 12.819101532118133,
+      "learning_rate": 3.8512396694214875e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0446,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.6875,
+      "step": 1280
+    },
+    {
+      "epoch": 1.5717791411042945,
+      "grad_norm": 12.392770855720375,
+      "learning_rate": 3.8479338842975207e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.470703125,
+      "logps/chosen": -432.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0467,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -8.6875,
+      "step": 1281
+    },
+    {
+      "epoch": 1.5730061349693252,
+      "grad_norm": 10.597792635666673,
+      "learning_rate": 3.8446280991735533e-07,
+      "logits/chosen": -0.04541015625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0369,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9140625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.375,
+      "step": 1282
+    },
+    {
+      "epoch": 1.5742331288343558,
+      "grad_norm": 13.234069958439528,
+      "learning_rate": 3.8413223140495865e-07,
+      "logits/chosen": -0.09521484375,
+      "logits/rejected": -0.2275390625,
+      "logps/chosen": -344.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0541,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.08251953125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.8125,
+      "step": 1283
+    },
+    {
+      "epoch": 1.5754601226993865,
+      "grad_norm": 21.414790419464424,
+      "learning_rate": 3.8380165289256197e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0854,
+      "rewards/accuracies": 0.953125,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.1875,
+      "step": 1284
+    },
+    {
+      "epoch": 1.5766871165644172,
+      "grad_norm": 14.621409356775658,
+      "learning_rate": 3.834710743801653e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0331,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.125,
+      "step": 1285
+    },
+    {
+      "epoch": 1.5779141104294478,
+      "grad_norm": 11.226540370477307,
+      "learning_rate": 3.831404958677686e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0366,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.318359375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.0625,
+      "step": 1286
+    },
+    {
+      "epoch": 1.5791411042944785,
+      "grad_norm": 16.039598977424404,
+      "learning_rate": 3.8280991735537187e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -318.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0613,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": -0.1640625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -9.3125,
+      "step": 1287
+    },
+    {
+      "epoch": 1.5803680981595092,
+      "grad_norm": 12.836865540318971,
+      "learning_rate": 3.824793388429752e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0505,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.65625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.0,
+      "step": 1288
+    },
+    {
+      "epoch": 1.5815950920245399,
+      "grad_norm": 7.945536104012517,
+      "learning_rate": 3.821487603305785e-07,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -324.0,
+      "logps/rejected": -318.0,
+      "loss": 0.043,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.072265625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -9.25,
+      "step": 1289
+    },
+    {
+      "epoch": 1.5828220858895705,
+      "grad_norm": 9.269509711129304,
+      "learning_rate": 3.818181818181818e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0298,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1533203125,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -10.125,
+      "step": 1290
+    },
+    {
+      "epoch": 1.5840490797546012,
+      "grad_norm": 12.776471775676947,
+      "learning_rate": 3.814876033057851e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -346.0,
+      "loss": 0.042,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.28125,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.3125,
+      "step": 1291
+    },
+    {
+      "epoch": 1.5852760736196319,
+      "grad_norm": 6.710032759787997,
+      "learning_rate": 3.811570247933884e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0285,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.412109375,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.875,
+      "step": 1292
+    },
+    {
+      "epoch": 1.5865030674846625,
+      "grad_norm": 9.554015835107709,
+      "learning_rate": 3.808264462809917e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0361,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.447265625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.0625,
+      "step": 1293
+    },
+    {
+      "epoch": 1.5877300613496934,
+      "grad_norm": 7.6540954238448995,
+      "learning_rate": 3.8049586776859504e-07,
+      "logits/chosen": -0.2001953125,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0283,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.4296875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.9375,
+      "step": 1294
+    },
+    {
+      "epoch": 1.588957055214724,
+      "grad_norm": 10.249402612807254,
+      "learning_rate": 3.8016528925619836e-07,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0528,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.025390625,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -9.1875,
+      "step": 1295
+    },
+    {
+      "epoch": 1.5901840490797547,
+      "grad_norm": 21.386881079863016,
+      "learning_rate": 3.798347107438016e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -374.0,
+      "loss": 0.1043,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": -0.474609375,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -9.25,
+      "step": 1296
+    },
+    {
+      "epoch": 1.5914110429447854,
+      "grad_norm": 13.13883607900759,
+      "learning_rate": 3.7950413223140494e-07,
+      "logits/chosen": -0.1513671875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0505,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.271484375,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -8.0625,
+      "step": 1297
+    },
+    {
+      "epoch": 1.592638036809816,
+      "grad_norm": 14.148949473846967,
+      "learning_rate": 3.7917355371900826e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0601,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.73828125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.1875,
+      "step": 1298
+    },
+    {
+      "epoch": 1.5938650306748468,
+      "grad_norm": 12.160973266195128,
+      "learning_rate": 3.788429752066116e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0443,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.2353515625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.875,
+      "step": 1299
+    },
+    {
+      "epoch": 1.5950920245398774,
+      "grad_norm": 11.389665455717337,
+      "learning_rate": 3.7851239669421484e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0408,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5546875,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.0,
+      "step": 1300
+    },
+    {
+      "epoch": 1.596319018404908,
+      "grad_norm": 12.77732004217483,
+      "learning_rate": 3.7818181818181816e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0424,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1162109375,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.8125,
+      "step": 1301
+    },
+    {
+      "epoch": 1.5975460122699388,
+      "grad_norm": 12.748510423611386,
+      "learning_rate": 3.778512396694215e-07,
+      "logits/chosen": -0.359375,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0616,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.0,
+      "step": 1302
+    },
+    {
+      "epoch": 1.5987730061349694,
+      "grad_norm": 8.32794148442266,
+      "learning_rate": 3.775206611570248e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0402,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.89453125,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.3125,
+      "step": 1303
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 6.529193782585084,
+      "learning_rate": 3.7719008264462806e-07,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0274,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.84375,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.125,
+      "step": 1304
+    },
+    {
+      "epoch": 1.6012269938650308,
+      "grad_norm": 9.816683371253708,
+      "learning_rate": 3.768595041322314e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -342.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0524,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.4375,
+      "step": 1305
+    },
+    {
+      "epoch": 1.6024539877300614,
+      "grad_norm": 8.33652199166503,
+      "learning_rate": 3.765289256198347e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0333,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.8125,
+      "step": 1306
+    },
+    {
+      "epoch": 1.603680981595092,
+      "grad_norm": 5.703980454598279,
+      "learning_rate": 3.76198347107438e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0237,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.1875,
+      "step": 1307
+    },
+    {
+      "epoch": 1.6049079754601228,
+      "grad_norm": 7.658766807333126,
+      "learning_rate": 3.7586776859504133e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0286,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.9140625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.4375,
+      "step": 1308
+    },
+    {
+      "epoch": 1.6061349693251534,
+      "grad_norm": 9.415605561419792,
+      "learning_rate": 3.755371900826446e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0372,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.50390625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.8125,
+      "step": 1309
+    },
+    {
+      "epoch": 1.607361963190184,
+      "grad_norm": 8.93902350488852,
+      "learning_rate": 3.752066115702479e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -330.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0376,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.375,
+      "step": 1310
+    },
+    {
+      "epoch": 1.6085889570552148,
+      "grad_norm": 12.399527167412383,
+      "learning_rate": 3.7487603305785123e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0605,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.5703125,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -7.59375,
+      "step": 1311
+    },
+    {
+      "epoch": 1.6098159509202454,
+      "grad_norm": 9.954372711363225,
+      "learning_rate": 3.7454545454545454e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0291,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.25,
+      "step": 1312
+    },
+    {
+      "epoch": 1.611042944785276,
+      "grad_norm": 8.770279110550883,
+      "learning_rate": 3.742148760330578e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0342,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.57421875,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.4375,
+      "step": 1313
+    },
+    {
+      "epoch": 1.6122699386503068,
+      "grad_norm": 8.372096906172947,
+      "learning_rate": 3.7388429752066113e-07,
+      "logits/chosen": -0.375,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0239,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.1875,
+      "step": 1314
+    },
+    {
+      "epoch": 1.6134969325153374,
+      "grad_norm": 13.050858951846532,
+      "learning_rate": 3.7355371900826445e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -446.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0497,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9375,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.5625,
+      "step": 1315
+    },
+    {
+      "epoch": 1.614723926380368,
+      "grad_norm": 11.087298553686718,
+      "learning_rate": 3.7322314049586776e-07,
+      "logits/chosen": -0.1103515625,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0421,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.423828125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.25,
+      "step": 1316
+    },
+    {
+      "epoch": 1.6159509202453988,
+      "grad_norm": 7.935377399082722,
+      "learning_rate": 3.728925619834711e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0349,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.5078125,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.3125,
+      "step": 1317
+    },
+    {
+      "epoch": 1.6171779141104294,
+      "grad_norm": 9.315066544333694,
+      "learning_rate": 3.7256198347107435e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0259,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.431640625,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.0,
+      "step": 1318
+    },
+    {
+      "epoch": 1.61840490797546,
+      "grad_norm": 10.045649594522278,
+      "learning_rate": 3.7223140495867766e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0314,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.8125,
+      "step": 1319
+    },
+    {
+      "epoch": 1.6196319018404908,
+      "grad_norm": 15.320474747736862,
+      "learning_rate": 3.71900826446281e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -378.0,
+      "loss": 0.048,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.3125,
+      "step": 1320
+    },
+    {
+      "epoch": 1.6208588957055214,
+      "grad_norm": 24.6231369547886,
+      "learning_rate": 3.715702479338843e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0705,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.94921875,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.8125,
+      "step": 1321
+    },
+    {
+      "epoch": 1.622085889570552,
+      "grad_norm": 10.301826595324812,
+      "learning_rate": 3.7123966942148756e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0333,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.474609375,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.75,
+      "step": 1322
+    },
+    {
+      "epoch": 1.6233128834355828,
+      "grad_norm": 5.800881401321108,
+      "learning_rate": 3.709090909090909e-07,
+      "logits/chosen": -0.09814453125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -330.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0247,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.66796875,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.6875,
+      "step": 1323
+    },
+    {
+      "epoch": 1.6245398773006134,
+      "grad_norm": 12.535973391030188,
+      "learning_rate": 3.705785123966942e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0588,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.8203125,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.6875,
+      "step": 1324
+    },
+    {
+      "epoch": 1.6257668711656441,
+      "grad_norm": 12.562789704449548,
+      "learning_rate": 3.702479338842975e-07,
+      "logits/chosen": -0.1357421875,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0523,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.353515625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.8125,
+      "step": 1325
+    },
+    {
+      "epoch": 1.6269938650306748,
+      "grad_norm": 7.829908833342415,
+      "learning_rate": 3.6991735537190083e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0269,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.4375,
+      "step": 1326
+    },
+    {
+      "epoch": 1.6282208588957054,
+      "grad_norm": 20.36070349744939,
+      "learning_rate": 3.695867768595041e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0663,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.1201171875,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.6875,
+      "step": 1327
+    },
+    {
+      "epoch": 1.6294478527607361,
+      "grad_norm": 15.025096253851128,
+      "learning_rate": 3.692561983471074e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0478,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.54296875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.25,
+      "step": 1328
+    },
+    {
+      "epoch": 1.6306748466257668,
+      "grad_norm": 11.82653566926932,
+      "learning_rate": 3.6892561983471073e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.30859375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.6875,
+      "step": 1329
+    },
+    {
+      "epoch": 1.6319018404907975,
+      "grad_norm": 10.863416901691561,
+      "learning_rate": 3.6859504132231405e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0477,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.21484375,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -9.0,
+      "step": 1330
+    },
+    {
+      "epoch": 1.6331288343558281,
+      "grad_norm": 10.689179527812403,
+      "learning_rate": 3.682644628099173e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -358.0,
+      "loss": 0.034,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.1455078125,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.875,
+      "step": 1331
+    },
+    {
+      "epoch": 1.6343558282208588,
+      "grad_norm": 10.19931808828623,
+      "learning_rate": 3.6793388429752063e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -364.0,
+      "loss": 0.031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.337890625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.1875,
+      "step": 1332
+    },
+    {
+      "epoch": 1.6355828220858895,
+      "grad_norm": 16.496319171690203,
+      "learning_rate": 3.6760330578512395e-07,
+      "logits/chosen": -0.341796875,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0631,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.30078125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.0625,
+      "step": 1333
+    },
+    {
+      "epoch": 1.6368098159509201,
+      "grad_norm": 10.236653831572811,
+      "learning_rate": 3.6727272727272727e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0362,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.98828125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.875,
+      "step": 1334
+    },
+    {
+      "epoch": 1.6380368098159508,
+      "grad_norm": 20.7027432046174,
+      "learning_rate": 3.669421487603306e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0518,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.3125,
+      "step": 1335
+    },
+    {
+      "epoch": 1.6392638036809815,
+      "grad_norm": 14.000221716050415,
+      "learning_rate": 3.6661157024793385e-07,
+      "logits/chosen": -0.41796875,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0466,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.064453125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.5,
+      "step": 1336
+    },
+    {
+      "epoch": 1.6404907975460121,
+      "grad_norm": 5.356677692115597,
+      "learning_rate": 3.6628099173553717e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0177,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.0,
+      "step": 1337
+    },
+    {
+      "epoch": 1.6417177914110428,
+      "grad_norm": 9.76155867846443,
+      "learning_rate": 3.659504132231405e-07,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -384.0,
+      "loss": 0.039,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.625,
+      "step": 1338
+    },
+    {
+      "epoch": 1.6429447852760735,
+      "grad_norm": 12.539356287206903,
+      "learning_rate": 3.656198347107438e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -322.0,
+      "loss": 0.0437,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.90625,
+      "step": 1339
+    },
+    {
+      "epoch": 1.6441717791411041,
+      "grad_norm": 18.075978090516806,
+      "learning_rate": 3.6528925619834707e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0864,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 0.83203125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.125,
+      "step": 1340
+    },
+    {
+      "epoch": 1.645398773006135,
+      "grad_norm": 7.930691006707937,
+      "learning_rate": 3.649586776859504e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -434.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0285,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.0625,
+      "step": 1341
+    },
+    {
+      "epoch": 1.6466257668711657,
+      "grad_norm": 7.577446727011239,
+      "learning_rate": 3.646280991735537e-07,
+      "logits/chosen": -0.31640625,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0396,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.92578125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.4375,
+      "step": 1342
+    },
+    {
+      "epoch": 1.6478527607361964,
+      "grad_norm": 33.752974596436516,
+      "learning_rate": 3.64297520661157e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -382.0,
+      "loss": 0.1004,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.373046875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.6875,
+      "step": 1343
+    },
+    {
+      "epoch": 1.649079754601227,
+      "grad_norm": 5.348550645984265,
+      "learning_rate": 3.639669421487603e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0171,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -8.75,
+      "step": 1344
+    },
+    {
+      "epoch": 1.6503067484662577,
+      "grad_norm": 8.28491808447824,
+      "learning_rate": 3.636363636363636e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0326,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7890625,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.90625,
+      "step": 1345
+    },
+    {
+      "epoch": 1.6515337423312884,
+      "grad_norm": 18.578868213165027,
+      "learning_rate": 3.633057851239669e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0489,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.4453125,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.3125,
+      "step": 1346
+    },
+    {
+      "epoch": 1.652760736196319,
+      "grad_norm": 19.63908403650041,
+      "learning_rate": 3.6297520661157024e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0988,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.4375,
+      "step": 1347
+    },
+    {
+      "epoch": 1.6539877300613497,
+      "grad_norm": 12.572025294012443,
+      "learning_rate": 3.6264462809917356e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -338.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0535,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.2099609375,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.875,
+      "step": 1348
+    },
+    {
+      "epoch": 1.6552147239263804,
+      "grad_norm": 8.973003562840935,
+      "learning_rate": 3.623140495867768e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0377,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.203125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.75,
+      "step": 1349
+    },
+    {
+      "epoch": 1.656441717791411,
+      "grad_norm": 15.006464506327205,
+      "learning_rate": 3.6198347107438014e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -412.0,
+      "loss": 0.05,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.125,
+      "step": 1350
+    },
+    {
+      "epoch": 1.6576687116564417,
+      "grad_norm": 17.74819569728375,
+      "learning_rate": 3.6165289256198346e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0841,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.486328125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.375,
+      "step": 1351
+    },
+    {
+      "epoch": 1.6588957055214724,
+      "grad_norm": 9.171226382603257,
+      "learning_rate": 3.613223140495868e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.470703125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0313,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.4296875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.8125,
+      "step": 1352
+    },
+    {
+      "epoch": 1.660122699386503,
+      "grad_norm": 8.31135298357435,
+      "learning_rate": 3.6099173553719004e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -396.0,
+      "loss": 0.039,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.017578125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -9.125,
+      "step": 1353
+    },
+    {
+      "epoch": 1.6613496932515337,
+      "grad_norm": 9.164109948648726,
+      "learning_rate": 3.6066115702479336e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0391,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.12451171875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.9375,
+      "step": 1354
+    },
+    {
+      "epoch": 1.6625766871165644,
+      "grad_norm": 10.936704965558942,
+      "learning_rate": 3.603305785123967e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0476,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.052978515625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.8125,
+      "step": 1355
+    },
+    {
+      "epoch": 1.6638036809815953,
+      "grad_norm": 13.709468392527326,
+      "learning_rate": 3.6e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.51953125,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -10.1875,
+      "step": 1356
+    },
+    {
+      "epoch": 1.665030674846626,
+      "grad_norm": 9.016472764379131,
+      "learning_rate": 3.596694214876033e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0216,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0673828125,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -10.25,
+      "step": 1357
+    },
+    {
+      "epoch": 1.6662576687116566,
+      "grad_norm": 20.157009863407833,
+      "learning_rate": 3.593388429752066e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0805,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.1181640625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.75,
+      "step": 1358
+    },
+    {
+      "epoch": 1.6674846625766873,
+      "grad_norm": 15.249365615978206,
+      "learning_rate": 3.590082644628099e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0406,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.39453125,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -10.1875,
+      "step": 1359
+    },
+    {
+      "epoch": 1.668711656441718,
+      "grad_norm": 16.208561543286084,
+      "learning_rate": 3.586776859504132e-07,
+      "logits/chosen": -0.45703125,
+      "logits/rejected": -0.69140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0488,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.384765625,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -10.375,
+      "step": 1360
+    },
+    {
+      "epoch": 1.6699386503067486,
+      "grad_norm": 10.803630772069788,
+      "learning_rate": 3.5834710743801653e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0402,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.255859375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.5,
+      "step": 1361
+    },
+    {
+      "epoch": 1.6711656441717793,
+      "grad_norm": 14.627959585498674,
+      "learning_rate": 3.580165289256198e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0524,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.361328125,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.875,
+      "step": 1362
+    },
+    {
+      "epoch": 1.67239263803681,
+      "grad_norm": 4.581091718122957,
+      "learning_rate": 3.576859504132231e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0148,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0791015625,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -10.5,
+      "step": 1363
+    },
+    {
+      "epoch": 1.6736196319018406,
+      "grad_norm": 8.359592521972438,
+      "learning_rate": 3.5735537190082643e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0237,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.515625,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.75,
+      "step": 1364
+    },
+    {
+      "epoch": 1.6748466257668713,
+      "grad_norm": 16.525640852041033,
+      "learning_rate": 3.5702479338842975e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0676,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.65625,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.5625,
+      "step": 1365
+    },
+    {
+      "epoch": 1.676073619631902,
+      "grad_norm": 7.921394403782825,
+      "learning_rate": 3.5669421487603307e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0262,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.83984375,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.3125,
+      "step": 1366
+    },
+    {
+      "epoch": 1.6773006134969326,
+      "grad_norm": 12.021872239150174,
+      "learning_rate": 3.5636363636363633e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.7734375,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.8125,
+      "step": 1367
+    },
+    {
+      "epoch": 1.6785276073619633,
+      "grad_norm": 18.310359770819986,
+      "learning_rate": 3.5603305785123965e-07,
+      "logits/chosen": -0.376953125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -378.0,
+      "loss": 0.057,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1357421875,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.9375,
+      "step": 1368
+    },
+    {
+      "epoch": 1.679754601226994,
+      "grad_norm": 16.65697818828575,
+      "learning_rate": 3.5570247933884297e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -380.0,
+      "loss": 0.079,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.0218505859375,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.3125,
+      "step": 1369
+    },
+    {
+      "epoch": 1.6809815950920246,
+      "grad_norm": 11.204270470130796,
+      "learning_rate": 3.553719008264463e-07,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0306,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.75,
+      "step": 1370
+    },
+    {
+      "epoch": 1.6822085889570553,
+      "grad_norm": 15.13195763037223,
+      "learning_rate": 3.5504132231404955e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -464.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0514,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -9.5,
+      "step": 1371
+    },
+    {
+      "epoch": 1.683435582822086,
+      "grad_norm": 8.16317454264987,
+      "learning_rate": 3.5471074380165287e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0346,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.4375,
+      "step": 1372
+    },
+    {
+      "epoch": 1.6846625766871166,
+      "grad_norm": 15.473022951774299,
+      "learning_rate": 3.543801652892562e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -394.0,
+      "loss": 0.05,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7265625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.0625,
+      "step": 1373
+    },
+    {
+      "epoch": 1.6858895705521473,
+      "grad_norm": 13.313790214938209,
+      "learning_rate": 3.540495867768595e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0592,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.3125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.6875,
+      "step": 1374
+    },
+    {
+      "epoch": 1.687116564417178,
+      "grad_norm": 15.234502433567197,
+      "learning_rate": 3.537190082644628e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -326.0,
+      "logps/rejected": -342.0,
+      "loss": 0.065,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.498046875,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.625,
+      "step": 1375
+    },
+    {
+      "epoch": 1.6883435582822086,
+      "grad_norm": 7.451858103319708,
+      "learning_rate": 3.533884297520661e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0294,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.375,
+      "step": 1376
+    },
+    {
+      "epoch": 1.6895705521472393,
+      "grad_norm": 11.164973482876846,
+      "learning_rate": 3.530578512396694e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0356,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -9.5625,
+      "step": 1377
+    },
+    {
+      "epoch": 1.69079754601227,
+      "grad_norm": 7.087310188668901,
+      "learning_rate": 3.527272727272727e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0308,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.193359375,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.5625,
+      "step": 1378
+    },
+    {
+      "epoch": 1.6920245398773006,
+      "grad_norm": 9.503069803683369,
+      "learning_rate": 3.5239669421487604e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0483,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.96875,
+      "step": 1379
+    },
+    {
+      "epoch": 1.6932515337423313,
+      "grad_norm": 6.484416144875946,
+      "learning_rate": 3.520661157024793e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0227,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.3125,
+      "step": 1380
+    },
+    {
+      "epoch": 1.694478527607362,
+      "grad_norm": 12.461259591914391,
+      "learning_rate": 3.517355371900826e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.1767578125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0506,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.482421875,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.9375,
+      "step": 1381
+    },
+    {
+      "epoch": 1.6957055214723926,
+      "grad_norm": 12.727643966429453,
+      "learning_rate": 3.5140495867768594e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0412,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2294921875,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.625,
+      "step": 1382
+    },
+    {
+      "epoch": 1.6969325153374233,
+      "grad_norm": 6.275694798930385,
+      "learning_rate": 3.5107438016528925e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0245,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.5625,
+      "step": 1383
+    },
+    {
+      "epoch": 1.698159509202454,
+      "grad_norm": 13.180464183457763,
+      "learning_rate": 3.5074380165289257e-07,
+      "logits/chosen": -0.095703125,
+      "logits/rejected": -0.21484375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0482,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.51171875,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.0625,
+      "step": 1384
+    },
+    {
+      "epoch": 1.6993865030674846,
+      "grad_norm": 13.726588628334937,
+      "learning_rate": 3.5041322314049584e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -326.0,
+      "logps/rejected": -326.0,
+      "loss": 0.058,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.68359375,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.25,
+      "step": 1385
+    },
+    {
+      "epoch": 1.7006134969325153,
+      "grad_norm": 9.154420481281178,
+      "learning_rate": 3.5008264462809915e-07,
+      "logits/chosen": -0.359375,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0368,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.0,
+      "step": 1386
+    },
+    {
+      "epoch": 1.701840490797546,
+      "grad_norm": 11.579193730406892,
+      "learning_rate": 3.4975206611570247e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0428,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.0,
+      "step": 1387
+    },
+    {
+      "epoch": 1.7030674846625766,
+      "grad_norm": 9.159728350094172,
+      "learning_rate": 3.494214876033058e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0348,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.390625,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -8.1875,
+      "step": 1388
+    },
+    {
+      "epoch": 1.7042944785276073,
+      "grad_norm": 10.976035161259295,
+      "learning_rate": 3.4909090909090905e-07,
+      "logits/chosen": -0.361328125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0444,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9140625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.9375,
+      "step": 1389
+    },
+    {
+      "epoch": 1.705521472392638,
+      "grad_norm": 15.764967882572376,
+      "learning_rate": 3.4876033057851237e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0452,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.58984375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.125,
+      "step": 1390
+    },
+    {
+      "epoch": 1.7067484662576686,
+      "grad_norm": 7.403388918790972,
+      "learning_rate": 3.484297520661157e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0356,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.375,
+      "step": 1391
+    },
+    {
+      "epoch": 1.7079754601226993,
+      "grad_norm": 10.896245766912692,
+      "learning_rate": 3.48099173553719e-07,
+      "logits/chosen": -0.10107421875,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.31640625,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -7.875,
+      "step": 1392
+    },
+    {
+      "epoch": 1.70920245398773,
+      "grad_norm": 11.366422205322182,
+      "learning_rate": 3.4776859504132227e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0456,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.0625,
+      "step": 1393
+    },
+    {
+      "epoch": 1.7104294478527606,
+      "grad_norm": 14.99148246159462,
+      "learning_rate": 3.474380165289256e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0628,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.8125,
+      "step": 1394
+    },
+    {
+      "epoch": 1.7116564417177913,
+      "grad_norm": 5.724911703113608,
+      "learning_rate": 3.471074380165289e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0234,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.25,
+      "step": 1395
+    },
+    {
+      "epoch": 1.712883435582822,
+      "grad_norm": 7.299931111351364,
+      "learning_rate": 3.467768595041322e-07,
+      "logits/chosen": -0.0869140625,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0281,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.953125,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.125,
+      "step": 1396
+    },
+    {
+      "epoch": 1.7141104294478526,
+      "grad_norm": 7.734982226412597,
+      "learning_rate": 3.4644628099173554e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0299,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.1875,
+      "step": 1397
+    },
+    {
+      "epoch": 1.7153374233128833,
+      "grad_norm": 12.433607504676552,
+      "learning_rate": 3.461157024793388e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0423,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.6875,
+      "step": 1398
+    },
+    {
+      "epoch": 1.716564417177914,
+      "grad_norm": 10.213873529131746,
+      "learning_rate": 3.457851239669421e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0372,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.494140625,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.625,
+      "step": 1399
+    },
+    {
+      "epoch": 1.7177914110429446,
+      "grad_norm": 8.219886818141926,
+      "learning_rate": 3.4545454545454544e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -340.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0241,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.025146484375,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.8125,
+      "step": 1400
+    },
+    {
+      "epoch": 1.7190184049079753,
+      "grad_norm": 8.611806636266707,
+      "learning_rate": 3.4512396694214876e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -444.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0354,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.25,
+      "step": 1401
+    },
+    {
+      "epoch": 1.720245398773006,
+      "grad_norm": 9.897114375609455,
+      "learning_rate": 3.44793388429752e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0394,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.578125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.9375,
+      "step": 1402
+    },
+    {
+      "epoch": 1.7214723926380369,
+      "grad_norm": 8.741002963663481,
+      "learning_rate": 3.4446280991735534e-07,
+      "logits/chosen": -0.1572265625,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0278,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.46484375,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.8125,
+      "step": 1403
+    },
+    {
+      "epoch": 1.7226993865030675,
+      "grad_norm": 10.367535092764449,
+      "learning_rate": 3.4413223140495866e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -334.0,
+      "logps/rejected": -308.0,
+      "loss": 0.0367,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.91796875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.6875,
+      "step": 1404
+    },
+    {
+      "epoch": 1.7239263803680982,
+      "grad_norm": 13.241758246923336,
+      "learning_rate": 3.43801652892562e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0467,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.55859375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.375,
+      "step": 1405
+    },
+    {
+      "epoch": 1.7251533742331289,
+      "grad_norm": 8.331965679319266,
+      "learning_rate": 3.434710743801653e-07,
+      "logits/chosen": -0.38671875,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -324.0,
+      "logps/rejected": -352.0,
+      "loss": 0.023,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.01019287109375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.9375,
+      "step": 1406
+    },
+    {
+      "epoch": 1.7263803680981595,
+      "grad_norm": 15.380429512755986,
+      "learning_rate": 3.4314049586776856e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -438.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0578,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.9375,
+      "step": 1407
+    },
+    {
+      "epoch": 1.7276073619631902,
+      "grad_norm": 11.962024064286275,
+      "learning_rate": 3.428099173553719e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -470.0,
+      "logps/rejected": -410.0,
+      "loss": 0.058,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.921875,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.1875,
+      "step": 1408
+    },
+    {
+      "epoch": 1.7288343558282209,
+      "grad_norm": 12.743570008480159,
+      "learning_rate": 3.424793388429752e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.330078125,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.25,
+      "step": 1409
+    },
+    {
+      "epoch": 1.7300613496932515,
+      "grad_norm": 10.912975195599188,
+      "learning_rate": 3.421487603305785e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0529,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.05126953125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -9.0,
+      "step": 1410
+    },
+    {
+      "epoch": 1.7312883435582822,
+      "grad_norm": 9.273924247986834,
+      "learning_rate": 3.418181818181818e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0371,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.0213623046875,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.375,
+      "step": 1411
+    },
+    {
+      "epoch": 1.7325153374233129,
+      "grad_norm": 16.100998085595243,
+      "learning_rate": 3.414876033057851e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0693,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.66015625,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.1875,
+      "step": 1412
+    },
+    {
+      "epoch": 1.7337423312883435,
+      "grad_norm": 12.422653573930514,
+      "learning_rate": 3.411570247933884e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0534,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.037353515625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.6875,
+      "step": 1413
+    },
+    {
+      "epoch": 1.7349693251533742,
+      "grad_norm": 8.035094608949246,
+      "learning_rate": 3.4082644628099173e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -432.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0358,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.5546875,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.0,
+      "step": 1414
+    },
+    {
+      "epoch": 1.7361963190184049,
+      "grad_norm": 15.842862209397916,
+      "learning_rate": 3.4049586776859505e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0571,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.265625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.25,
+      "step": 1415
+    },
+    {
+      "epoch": 1.7374233128834355,
+      "grad_norm": 12.823306849806624,
+      "learning_rate": 3.401652892561983e-07,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0456,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9921875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.0625,
+      "step": 1416
+    },
+    {
+      "epoch": 1.7386503067484662,
+      "grad_norm": 9.156727463682909,
+      "learning_rate": 3.3983471074380163e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0337,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5546875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.5,
+      "step": 1417
+    },
+    {
+      "epoch": 1.7398773006134969,
+      "grad_norm": 11.45003765016041,
+      "learning_rate": 3.3950413223140495e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0371,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.3125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.5625,
+      "step": 1418
+    },
+    {
+      "epoch": 1.7411042944785278,
+      "grad_norm": 12.920796689779438,
+      "learning_rate": 3.3917355371900827e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -340.0,
+      "logps/rejected": -318.0,
+      "loss": 0.0579,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.11572265625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.75,
+      "step": 1419
+    },
+    {
+      "epoch": 1.7423312883435584,
+      "grad_norm": 14.910027932916357,
+      "learning_rate": 3.3884297520661153e-07,
+      "logits/chosen": -0.396484375,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -396.0,
+      "loss": 0.039,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.57421875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.8125,
+      "step": 1420
+    },
+    {
+      "epoch": 1.743558282208589,
+      "grad_norm": 8.893525346196068,
+      "learning_rate": 3.3851239669421485e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.5,
+      "logps/chosen": -376.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0345,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.53515625,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.875,
+      "step": 1421
+    },
+    {
+      "epoch": 1.7447852760736198,
+      "grad_norm": 15.365028675388714,
+      "learning_rate": 3.3818181818181817e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0651,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.3046875,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.9375,
+      "step": 1422
+    },
+    {
+      "epoch": 1.7460122699386504,
+      "grad_norm": 9.21774061389196,
+      "learning_rate": 3.378512396694215e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0397,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.3828125,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -10.5625,
+      "step": 1423
+    },
+    {
+      "epoch": 1.747239263803681,
+      "grad_norm": 10.515642559799373,
+      "learning_rate": 3.375206611570248e-07,
+      "logits/chosen": -0.330078125,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0431,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.333984375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.625,
+      "step": 1424
+    },
+    {
+      "epoch": 1.7484662576687118,
+      "grad_norm": 10.498215748420622,
+      "learning_rate": 3.3719008264462807e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0385,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.431640625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.625,
+      "step": 1425
+    },
+    {
+      "epoch": 1.7496932515337424,
+      "grad_norm": 13.42107772359636,
+      "learning_rate": 3.368595041322314e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0665,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.427734375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.25,
+      "step": 1426
+    },
+    {
+      "epoch": 1.7509202453987731,
+      "grad_norm": 17.515697940964518,
+      "learning_rate": 3.365289256198347e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0599,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.5625,
+      "step": 1427
+    },
+    {
+      "epoch": 1.7521472392638038,
+      "grad_norm": 12.347321878732972,
+      "learning_rate": 3.36198347107438e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0456,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9296875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.3125,
+      "step": 1428
+    },
+    {
+      "epoch": 1.7533742331288344,
+      "grad_norm": 11.083184332644388,
+      "learning_rate": 3.358677685950413e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0441,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.59375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.0,
+      "step": 1429
+    },
+    {
+      "epoch": 1.7546012269938651,
+      "grad_norm": 7.888788470102411,
+      "learning_rate": 3.355371900826446e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0322,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.58203125,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.1875,
+      "step": 1430
+    },
+    {
+      "epoch": 1.7558282208588958,
+      "grad_norm": 10.25363004538014,
+      "learning_rate": 3.352066115702479e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0383,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.44921875,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.0625,
+      "step": 1431
+    },
+    {
+      "epoch": 1.7570552147239265,
+      "grad_norm": 5.057831122485827,
+      "learning_rate": 3.3487603305785124e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -334.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0153,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.25,
+      "step": 1432
+    },
+    {
+      "epoch": 1.7582822085889571,
+      "grad_norm": 17.51301080467881,
+      "learning_rate": 3.345454545454545e-07,
+      "logits/chosen": -0.10107421875,
+      "logits/rejected": -0.21484375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0887,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.13671875,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -8.0,
+      "step": 1433
+    },
+    {
+      "epoch": 1.7595092024539878,
+      "grad_norm": 11.377580450972712,
+      "learning_rate": 3.342148760330578e-07,
+      "logits/chosen": -0.126953125,
+      "logits/rejected": -0.2275390625,
+      "logps/chosen": -342.0,
+      "logps/rejected": -390.0,
+      "loss": 0.045,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.357421875,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.25,
+      "step": 1434
+    },
+    {
+      "epoch": 1.7607361963190185,
+      "grad_norm": 8.026386246513397,
+      "learning_rate": 3.3388429752066114e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -348.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0306,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.57421875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.375,
+      "step": 1435
+    },
+    {
+      "epoch": 1.7619631901840491,
+      "grad_norm": 5.610915011446518,
+      "learning_rate": 3.3355371900826446e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0205,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -8.875,
+      "step": 1436
+    },
+    {
+      "epoch": 1.7631901840490798,
+      "grad_norm": 12.950887212243945,
+      "learning_rate": 3.332231404958678e-07,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0445,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.1875,
+      "step": 1437
+    },
+    {
+      "epoch": 1.7644171779141105,
+      "grad_norm": 12.771531747113794,
+      "learning_rate": 3.3289256198347104e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0399,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.42578125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.4375,
+      "step": 1438
+    },
+    {
+      "epoch": 1.7656441717791411,
+      "grad_norm": 9.19835013594589,
+      "learning_rate": 3.3256198347107436e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0317,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.027099609375,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -8.5625,
+      "step": 1439
+    },
+    {
+      "epoch": 1.7668711656441718,
+      "grad_norm": 7.045378068044565,
+      "learning_rate": 3.322314049586777e-07,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -466.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0232,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.25,
+      "step": 1440
+    },
+    {
+      "epoch": 1.7680981595092025,
+      "grad_norm": 11.45812069017029,
+      "learning_rate": 3.31900826446281e-07,
+      "logits/chosen": -0.37890625,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0426,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.443359375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.625,
+      "step": 1441
+    },
+    {
+      "epoch": 1.7693251533742331,
+      "grad_norm": 15.329344086941836,
+      "learning_rate": 3.3157024793388426e-07,
+      "logits/chosen": -0.322265625,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -430.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0525,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.490234375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.125,
+      "step": 1442
+    },
+    {
+      "epoch": 1.7705521472392638,
+      "grad_norm": 8.1454122193256,
+      "learning_rate": 3.312396694214876e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0342,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9375,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.1875,
+      "step": 1443
+    },
+    {
+      "epoch": 1.7717791411042945,
+      "grad_norm": 18.4128990711308,
+      "learning_rate": 3.309090909090909e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0418,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.0625,
+      "step": 1444
+    },
+    {
+      "epoch": 1.7730061349693251,
+      "grad_norm": 10.577537024132393,
+      "learning_rate": 3.305785123966942e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -326.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0331,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1025390625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -9.0625,
+      "step": 1445
+    },
+    {
+      "epoch": 1.7742331288343558,
+      "grad_norm": 14.48140279792318,
+      "learning_rate": 3.3024793388429753e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0426,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.486328125,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -10.0,
+      "step": 1446
+    },
+    {
+      "epoch": 1.7754601226993865,
+      "grad_norm": 10.710617346705183,
+      "learning_rate": 3.299173553719008e-07,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0422,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1640625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.5625,
+      "step": 1447
+    },
+    {
+      "epoch": 1.7766871165644171,
+      "grad_norm": 13.124515407591634,
+      "learning_rate": 3.295867768595041e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -342.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0472,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.3359375,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -9.375,
+      "step": 1448
+    },
+    {
+      "epoch": 1.7779141104294478,
+      "grad_norm": 11.195883118200678,
+      "learning_rate": 3.2925619834710743e-07,
+      "logits/chosen": -0.10595703125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0415,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.2490234375,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -9.0,
+      "step": 1449
+    },
+    {
+      "epoch": 1.7791411042944785,
+      "grad_norm": 13.381809497018025,
+      "learning_rate": 3.2892561983471075e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.421875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.3125,
+      "step": 1450
+    },
+    {
+      "epoch": 1.7803680981595091,
+      "grad_norm": 9.53349976877088,
+      "learning_rate": 3.28595041322314e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0383,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.5390625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.3125,
+      "step": 1451
+    },
+    {
+      "epoch": 1.7815950920245398,
+      "grad_norm": 14.85880299692277,
+      "learning_rate": 3.2826446280991733e-07,
+      "logits/chosen": -0.373046875,
+      "logits/rejected": -0.625,
+      "logps/chosen": -454.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0233,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -9.75,
+      "step": 1452
+    },
+    {
+      "epoch": 1.7828220858895705,
+      "grad_norm": 7.4366621556306605,
+      "learning_rate": 3.2793388429752065e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0244,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.8125,
+      "step": 1453
+    },
+    {
+      "epoch": 1.7840490797546011,
+      "grad_norm": 14.1195355778993,
+      "learning_rate": 3.2760330578512396e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0325,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.27734375,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.9375,
+      "step": 1454
+    },
+    {
+      "epoch": 1.7852760736196318,
+      "grad_norm": 9.184961337980399,
+      "learning_rate": 3.272727272727273e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0351,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.875,
+      "step": 1455
+    },
+    {
+      "epoch": 1.7865030674846625,
+      "grad_norm": 11.51120353006608,
+      "learning_rate": 3.2694214876033055e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0465,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.25390625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.375,
+      "step": 1456
+    },
+    {
+      "epoch": 1.7877300613496931,
+      "grad_norm": 10.80392052881998,
+      "learning_rate": 3.2661157024793386e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -420.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0384,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.455078125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.4375,
+      "step": 1457
+    },
+    {
+      "epoch": 1.7889570552147238,
+      "grad_norm": 19.370928970130752,
+      "learning_rate": 3.262809917355372e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.2421875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0589,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.04345703125,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.4375,
+      "step": 1458
+    },
+    {
+      "epoch": 1.7901840490797545,
+      "grad_norm": 8.14401503004669,
+      "learning_rate": 3.259504132231405e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0249,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.73828125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.6875,
+      "step": 1459
+    },
+    {
+      "epoch": 1.7914110429447851,
+      "grad_norm": 8.410071233667539,
+      "learning_rate": 3.2561983471074376e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0337,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.796875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.8125,
+      "step": 1460
+    },
+    {
+      "epoch": 1.7926380368098158,
+      "grad_norm": 17.319232365393777,
+      "learning_rate": 3.252892561983471e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0584,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8203125,
+      "rewards/margins": 11.125,
+      "rewards/rejected": -10.3125,
+      "step": 1461
+    },
+    {
+      "epoch": 1.7938650306748465,
+      "grad_norm": 11.658403751664542,
+      "learning_rate": 3.249586776859504e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -324.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0423,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.412109375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.125,
+      "step": 1462
+    },
+    {
+      "epoch": 1.7950920245398772,
+      "grad_norm": 6.975630646192431,
+      "learning_rate": 3.246280991735537e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0238,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.65625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.875,
+      "step": 1463
+    },
+    {
+      "epoch": 1.7963190184049078,
+      "grad_norm": 13.732283507159519,
+      "learning_rate": 3.2429752066115704e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0461,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.294921875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.5,
+      "step": 1464
+    },
+    {
+      "epoch": 1.7975460122699385,
+      "grad_norm": 9.668785840082787,
+      "learning_rate": 3.239669421487603e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0301,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.6171875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.625,
+      "step": 1465
+    },
+    {
+      "epoch": 1.7987730061349694,
+      "grad_norm": 5.445268106615218,
+      "learning_rate": 3.236363636363636e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0153,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1328125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.8125,
+      "step": 1466
+    },
+    {
+      "epoch": 1.8,
+      "grad_norm": 8.337653458750603,
+      "learning_rate": 3.2330578512396694e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0287,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.625,
+      "step": 1467
+    },
+    {
+      "epoch": 1.8012269938650307,
+      "grad_norm": 14.119670175365039,
+      "learning_rate": 3.2297520661157025e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0381,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.0625,
+      "step": 1468
+    },
+    {
+      "epoch": 1.8024539877300614,
+      "grad_norm": 15.57267164353982,
+      "learning_rate": 3.226446280991735e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0642,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.32421875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.375,
+      "step": 1469
+    },
+    {
+      "epoch": 1.803680981595092,
+      "grad_norm": 13.78094405933175,
+      "learning_rate": 3.2231404958677684e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0781,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.022216796875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.8125,
+      "step": 1470
+    },
+    {
+      "epoch": 1.8049079754601227,
+      "grad_norm": 8.008772562359416,
+      "learning_rate": 3.2198347107438015e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -374.0,
+      "loss": 0.03,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.4375,
+      "step": 1471
+    },
+    {
+      "epoch": 1.8061349693251534,
+      "grad_norm": 15.98977849235284,
+      "learning_rate": 3.2165289256198347e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0514,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.75,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.125,
+      "step": 1472
+    },
+    {
+      "epoch": 1.807361963190184,
+      "grad_norm": 9.291346547833697,
+      "learning_rate": 3.2132231404958674e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -362.0,
+      "loss": 0.037,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.83984375,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.5,
+      "step": 1473
+    },
+    {
+      "epoch": 1.8085889570552147,
+      "grad_norm": 18.454994761707635,
+      "learning_rate": 3.2099173553719005e-07,
+      "logits/chosen": -0.10107421875,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0443,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.9609375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.0,
+      "step": 1474
+    },
+    {
+      "epoch": 1.8098159509202454,
+      "grad_norm": 9.644230934604456,
+      "learning_rate": 3.2066115702479337e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0343,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7265625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.0,
+      "step": 1475
+    },
+    {
+      "epoch": 1.811042944785276,
+      "grad_norm": 6.913662419393237,
+      "learning_rate": 3.203305785123967e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0234,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.3125,
+      "step": 1476
+    },
+    {
+      "epoch": 1.8122699386503067,
+      "grad_norm": 13.89342979775928,
+      "learning_rate": 3.2e-07,
+      "logits/chosen": -0.1396484375,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0585,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.419921875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.0,
+      "step": 1477
+    },
+    {
+      "epoch": 1.8134969325153374,
+      "grad_norm": 12.2206513449555,
+      "learning_rate": 3.1966942148760327e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0407,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4296875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -8.8125,
+      "step": 1478
+    },
+    {
+      "epoch": 1.814723926380368,
+      "grad_norm": 17.54721691614464,
+      "learning_rate": 3.193388429752066e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0468,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.88671875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.5,
+      "step": 1479
+    },
+    {
+      "epoch": 1.8159509202453987,
+      "grad_norm": 18.143441483594128,
+      "learning_rate": 3.190082644628099e-07,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -392.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0632,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.25,
+      "step": 1480
+    },
+    {
+      "epoch": 1.8171779141104294,
+      "grad_norm": 16.606911484964844,
+      "learning_rate": 3.186776859504132e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.384765625,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.5625,
+      "step": 1481
+    },
+    {
+      "epoch": 1.8184049079754603,
+      "grad_norm": 13.154293841882401,
+      "learning_rate": 3.183471074380165e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0766,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.5390625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.6875,
+      "step": 1482
+    },
+    {
+      "epoch": 1.819631901840491,
+      "grad_norm": 14.861363357929438,
+      "learning_rate": 3.180165289256198e-07,
+      "logits/chosen": -0.0830078125,
+      "logits/rejected": -0.2431640625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -316.0,
+      "loss": 0.0629,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.345703125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.6875,
+      "step": 1483
+    },
+    {
+      "epoch": 1.8208588957055216,
+      "grad_norm": 14.681356413204243,
+      "learning_rate": 3.176859504132231e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -348.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0461,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1376953125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -9.0,
+      "step": 1484
+    },
+    {
+      "epoch": 1.8220858895705523,
+      "grad_norm": 10.743142311605123,
+      "learning_rate": 3.1735537190082644e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0294,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.004547119140625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.3125,
+      "step": 1485
+    },
+    {
+      "epoch": 1.823312883435583,
+      "grad_norm": 15.212920486790656,
+      "learning_rate": 3.1702479338842976e-07,
+      "logits/chosen": -0.353515625,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0613,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.412109375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.625,
+      "step": 1486
+    },
+    {
+      "epoch": 1.8245398773006136,
+      "grad_norm": 7.09810262598987,
+      "learning_rate": 3.16694214876033e-07,
+      "logits/chosen": -0.40234375,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0222,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.875,
+      "step": 1487
+    },
+    {
+      "epoch": 1.8257668711656443,
+      "grad_norm": 13.16618442532596,
+      "learning_rate": 3.1636363636363634e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0301,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.734375,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.6875,
+      "step": 1488
+    },
+    {
+      "epoch": 1.826993865030675,
+      "grad_norm": 13.779672103672336,
+      "learning_rate": 3.1603305785123966e-07,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -380.0,
+      "loss": 0.051,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.796875,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -9.875,
+      "step": 1489
+    },
+    {
+      "epoch": 1.8282208588957056,
+      "grad_norm": 11.957973151160742,
+      "learning_rate": 3.15702479338843e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0395,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.84375,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.375,
+      "step": 1490
+    },
+    {
+      "epoch": 1.8294478527607363,
+      "grad_norm": 12.913847461598793,
+      "learning_rate": 3.1537190082644624e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -480.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0464,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -9.6875,
+      "step": 1491
+    },
+    {
+      "epoch": 1.830674846625767,
+      "grad_norm": 13.516942361236422,
+      "learning_rate": 3.1504132231404956e-07,
+      "logits/chosen": -0.412109375,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0473,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.126953125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.1875,
+      "step": 1492
+    },
+    {
+      "epoch": 1.8319018404907976,
+      "grad_norm": 9.02996772401457,
+      "learning_rate": 3.147107438016529e-07,
+      "logits/chosen": -0.365234375,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0384,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.357421875,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -9.5625,
+      "step": 1493
+    },
+    {
+      "epoch": 1.8331288343558283,
+      "grad_norm": 7.984291330585717,
+      "learning_rate": 3.143801652892562e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -438.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0342,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.318359375,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.6875,
+      "step": 1494
+    },
+    {
+      "epoch": 1.834355828220859,
+      "grad_norm": 12.136279210243362,
+      "learning_rate": 3.140495867768595e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0394,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.462890625,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.875,
+      "step": 1495
+    },
+    {
+      "epoch": 1.8355828220858896,
+      "grad_norm": 7.676908976526212,
+      "learning_rate": 3.137190082644628e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0219,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.80859375,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.0625,
+      "step": 1496
+    },
+    {
+      "epoch": 1.8368098159509203,
+      "grad_norm": 6.418224453171171,
+      "learning_rate": 3.133884297520661e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0244,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.9375,
+      "step": 1497
+    },
+    {
+      "epoch": 1.838036809815951,
+      "grad_norm": 14.435468659209993,
+      "learning_rate": 3.130578512396694e-07,
+      "logits/chosen": -0.37890625,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0507,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.392578125,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.0625,
+      "step": 1498
+    },
+    {
+      "epoch": 1.8392638036809816,
+      "grad_norm": 7.358594644326529,
+      "learning_rate": 3.1272727272727273e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0276,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.89453125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.0,
+      "step": 1499
+    },
+    {
+      "epoch": 1.8404907975460123,
+      "grad_norm": 12.861833790262308,
+      "learning_rate": 3.12396694214876e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0584,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -8.625,
+      "step": 1500
+    },
+    {
+      "epoch": 1.841717791411043,
+      "grad_norm": 9.440131410139509,
+      "learning_rate": 3.120661157024793e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -376.0,
+      "loss": 0.036,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.0,
+      "step": 1501
+    },
+    {
+      "epoch": 1.8429447852760736,
+      "grad_norm": 10.759450623282584,
+      "learning_rate": 3.1173553719008263e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0313,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.24609375,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -9.4375,
+      "step": 1502
+    },
+    {
+      "epoch": 1.8441717791411043,
+      "grad_norm": 13.44299298290354,
+      "learning_rate": 3.1140495867768595e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0429,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.375,
+      "step": 1503
+    },
+    {
+      "epoch": 1.845398773006135,
+      "grad_norm": 10.171235600186767,
+      "learning_rate": 3.1107438016528927e-07,
+      "logits/chosen": -0.322265625,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0353,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.125,
+      "step": 1504
+    },
+    {
+      "epoch": 1.8466257668711656,
+      "grad_norm": 10.219435819353786,
+      "learning_rate": 3.1074380165289253e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -398.0,
+      "loss": 0.04,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.220703125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.125,
+      "step": 1505
+    },
+    {
+      "epoch": 1.8478527607361963,
+      "grad_norm": 8.678347100798598,
+      "learning_rate": 3.1041322314049585e-07,
+      "logits/chosen": -0.248046875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0319,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.0145263671875,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.3125,
+      "step": 1506
+    },
+    {
+      "epoch": 1.849079754601227,
+      "grad_norm": 12.945184798310114,
+      "learning_rate": 3.1008264462809917e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0488,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.87890625,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.375,
+      "step": 1507
+    },
+    {
+      "epoch": 1.8503067484662576,
+      "grad_norm": 10.145841916730342,
+      "learning_rate": 3.097520661157025e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0347,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1806640625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.125,
+      "step": 1508
+    },
+    {
+      "epoch": 1.8515337423312883,
+      "grad_norm": 9.354504262490346,
+      "learning_rate": 3.0942148760330575e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0279,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.8125,
+      "step": 1509
+    },
+    {
+      "epoch": 1.852760736196319,
+      "grad_norm": 9.721886786793712,
+      "learning_rate": 3.0909090909090907e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0299,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.75,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.4375,
+      "step": 1510
+    },
+    {
+      "epoch": 1.8539877300613496,
+      "grad_norm": 6.801235608742616,
+      "learning_rate": 3.087603305785124e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0197,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -10.0625,
+      "step": 1511
+    },
+    {
+      "epoch": 1.8552147239263803,
+      "grad_norm": 9.356069818038527,
+      "learning_rate": 3.084297520661157e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0352,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -9.6875,
+      "step": 1512
+    },
+    {
+      "epoch": 1.856441717791411,
+      "grad_norm": 15.1733760873437,
+      "learning_rate": 3.08099173553719e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0447,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.337890625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.375,
+      "step": 1513
+    },
+    {
+      "epoch": 1.8576687116564417,
+      "grad_norm": 8.210480914006832,
+      "learning_rate": 3.077685950413223e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0348,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.671875,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -10.5625,
+      "step": 1514
+    },
+    {
+      "epoch": 1.8588957055214723,
+      "grad_norm": 15.498698060895602,
+      "learning_rate": 3.074380165289256e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0445,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -10.0,
+      "step": 1515
+    },
+    {
+      "epoch": 1.860122699386503,
+      "grad_norm": 11.441149191311155,
+      "learning_rate": 3.071074380165289e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0493,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.3671875,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -9.5625,
+      "step": 1516
+    },
+    {
+      "epoch": 1.8613496932515337,
+      "grad_norm": 6.593493238446448,
+      "learning_rate": 3.0677685950413224e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0262,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.84765625,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -9.9375,
+      "step": 1517
+    },
+    {
+      "epoch": 1.8625766871165643,
+      "grad_norm": 13.313579919939231,
+      "learning_rate": 3.064462809917355e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0479,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.21875,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -10.25,
+      "step": 1518
+    },
+    {
+      "epoch": 1.863803680981595,
+      "grad_norm": 18.06842961008156,
+      "learning_rate": 3.061157024793388e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0436,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.67578125,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -9.9375,
+      "step": 1519
+    },
+    {
+      "epoch": 1.8650306748466257,
+      "grad_norm": 6.182876979289249,
+      "learning_rate": 3.0578512396694214e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0203,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.026611328125,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -10.25,
+      "step": 1520
+    },
+    {
+      "epoch": 1.8662576687116563,
+      "grad_norm": 13.565119222091443,
+      "learning_rate": 3.0545454545454546e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.046630859375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.375,
+      "step": 1521
+    },
+    {
+      "epoch": 1.867484662576687,
+      "grad_norm": 11.55315965721378,
+      "learning_rate": 3.051239669421487e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0499,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.546875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.5,
+      "step": 1522
+    },
+    {
+      "epoch": 1.8687116564417177,
+      "grad_norm": 8.768625848422289,
+      "learning_rate": 3.0479338842975204e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0266,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.515625,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.75,
+      "step": 1523
+    },
+    {
+      "epoch": 1.8699386503067483,
+      "grad_norm": 17.46967243494328,
+      "learning_rate": 3.0446280991735536e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0648,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.47265625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.9375,
+      "step": 1524
+    },
+    {
+      "epoch": 1.871165644171779,
+      "grad_norm": 18.901627284827356,
+      "learning_rate": 3.041322314049587e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -10.0,
+      "step": 1525
+    },
+    {
+      "epoch": 1.8723926380368097,
+      "grad_norm": 15.46265759049967,
+      "learning_rate": 3.03801652892562e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0669,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.9375,
+      "step": 1526
+    },
+    {
+      "epoch": 1.8736196319018403,
+      "grad_norm": 15.793631466248941,
+      "learning_rate": 3.0347107438016526e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0551,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.125,
+      "step": 1527
+    },
+    {
+      "epoch": 1.874846625766871,
+      "grad_norm": 8.116076751113416,
+      "learning_rate": 3.031404958677686e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0452,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -9.5625,
+      "step": 1528
+    },
+    {
+      "epoch": 1.876073619631902,
+      "grad_norm": 7.632384073577,
+      "learning_rate": 3.028099173553719e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0225,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -10.1875,
+      "step": 1529
+    },
+    {
+      "epoch": 1.8773006134969326,
+      "grad_norm": 13.902306225751913,
+      "learning_rate": 3.024793388429752e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0604,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.408203125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.9375,
+      "step": 1530
+    },
+    {
+      "epoch": 1.8785276073619632,
+      "grad_norm": 15.95752061813539,
+      "learning_rate": 3.021487603305785e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0628,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": -0.201171875,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.5625,
+      "step": 1531
+    },
+    {
+      "epoch": 1.879754601226994,
+      "grad_norm": 13.945969752259758,
+      "learning_rate": 3.018181818181818e-07,
+      "logits/chosen": -0.3984375,
+      "logits/rejected": -0.57421875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -422.0,
+      "loss": 0.04,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -10.0625,
+      "step": 1532
+    },
+    {
+      "epoch": 1.8809815950920246,
+      "grad_norm": 9.516345960706976,
+      "learning_rate": 3.014876033057851e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0406,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.416015625,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.875,
+      "step": 1533
+    },
+    {
+      "epoch": 1.8822085889570552,
+      "grad_norm": 9.285807586877766,
+      "learning_rate": 3.0115702479338843e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0353,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1171875,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.8125,
+      "step": 1534
+    },
+    {
+      "epoch": 1.883435582822086,
+      "grad_norm": 16.084894756892087,
+      "learning_rate": 3.0082644628099175e-07,
+      "logits/chosen": -0.1513671875,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0633,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8828125,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.125,
+      "step": 1535
+    },
+    {
+      "epoch": 1.8846625766871166,
+      "grad_norm": 18.252488190346416,
+      "learning_rate": 3.00495867768595e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -450.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0635,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.25,
+      "step": 1536
+    },
+    {
+      "epoch": 1.8858895705521472,
+      "grad_norm": 6.3990194084358984,
+      "learning_rate": 3.0016528925619833e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0197,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5703125,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.75,
+      "step": 1537
+    },
+    {
+      "epoch": 1.887116564417178,
+      "grad_norm": 6.78081351935863,
+      "learning_rate": 2.9983471074380165e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0168,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.7890625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.375,
+      "step": 1538
+    },
+    {
+      "epoch": 1.8883435582822086,
+      "grad_norm": 8.405892255575104,
+      "learning_rate": 2.9950413223140496e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0352,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.8125,
+      "step": 1539
+    },
+    {
+      "epoch": 1.8895705521472392,
+      "grad_norm": 10.388061012954786,
+      "learning_rate": 2.9917355371900823e-07,
+      "logits/chosen": -0.39453125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -442.0,
+      "loss": 0.04,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.5,
+      "step": 1540
+    },
+    {
+      "epoch": 1.89079754601227,
+      "grad_norm": 11.04941546644624,
+      "learning_rate": 2.9884297520661155e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -380.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.50390625,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.5,
+      "step": 1541
+    },
+    {
+      "epoch": 1.8920245398773006,
+      "grad_norm": 9.143600142098991,
+      "learning_rate": 2.9851239669421486e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0244,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -9.8125,
+      "step": 1542
+    },
+    {
+      "epoch": 1.8932515337423312,
+      "grad_norm": 6.647581423330869,
+      "learning_rate": 2.981818181818182e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0238,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.61328125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.875,
+      "step": 1543
+    },
+    {
+      "epoch": 1.8944785276073621,
+      "grad_norm": 8.984495028391082,
+      "learning_rate": 2.978512396694215e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -426.0,
+      "loss": 0.034,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.78515625,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.75,
+      "step": 1544
+    },
+    {
+      "epoch": 1.8957055214723928,
+      "grad_norm": 8.685599093816752,
+      "learning_rate": 2.9752066115702476e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0309,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.040771484375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.9375,
+      "step": 1545
+    },
+    {
+      "epoch": 1.8969325153374235,
+      "grad_norm": 17.402219988439615,
+      "learning_rate": 2.971900826446281e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.46875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.375,
+      "step": 1546
+    },
+    {
+      "epoch": 1.8981595092024541,
+      "grad_norm": 14.691890087949803,
+      "learning_rate": 2.968595041322314e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -484.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0367,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.91796875,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -10.4375,
+      "step": 1547
+    },
+    {
+      "epoch": 1.8993865030674848,
+      "grad_norm": 28.990198997957158,
+      "learning_rate": 2.965289256198347e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0523,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.369140625,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.4375,
+      "step": 1548
+    },
+    {
+      "epoch": 1.9006134969325155,
+      "grad_norm": 9.555007462448875,
+      "learning_rate": 2.96198347107438e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.322265625,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.0625,
+      "step": 1549
+    },
+    {
+      "epoch": 1.9018404907975461,
+      "grad_norm": 14.483826666374615,
+      "learning_rate": 2.958677685950413e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.423828125,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -10.125,
+      "step": 1550
+    },
+    {
+      "epoch": 1.9030674846625768,
+      "grad_norm": 13.761590699277777,
+      "learning_rate": 2.955371900826446e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -430.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0492,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.01031494140625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.625,
+      "step": 1551
+    },
+    {
+      "epoch": 1.9042944785276075,
+      "grad_norm": 13.384753485980882,
+      "learning_rate": 2.9520661157024793e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0606,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.15234375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.5,
+      "step": 1552
+    },
+    {
+      "epoch": 1.9055214723926381,
+      "grad_norm": 14.733975887876666,
+      "learning_rate": 2.9487603305785125e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0354,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.3046875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.5,
+      "step": 1553
+    },
+    {
+      "epoch": 1.9067484662576688,
+      "grad_norm": 13.95223783378209,
+      "learning_rate": 2.945454545454545e-07,
+      "logits/chosen": -0.369140625,
+      "logits/rejected": -0.6328125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0454,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.4140625,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.875,
+      "step": 1554
+    },
+    {
+      "epoch": 1.9079754601226995,
+      "grad_norm": 11.03554196573526,
+      "learning_rate": 2.9421487603305783e-07,
+      "logits/chosen": -0.390625,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0363,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.41015625,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.8125,
+      "step": 1555
+    },
+    {
+      "epoch": 1.9092024539877301,
+      "grad_norm": 9.08340310647596,
+      "learning_rate": 2.9388429752066115e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0364,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.64453125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.3125,
+      "step": 1556
+    },
+    {
+      "epoch": 1.9104294478527608,
+      "grad_norm": 12.6525315766818,
+      "learning_rate": 2.9355371900826447e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0477,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.330078125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.0,
+      "step": 1557
+    },
+    {
+      "epoch": 1.9116564417177915,
+      "grad_norm": 9.233008765341841,
+      "learning_rate": 2.9322314049586773e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0246,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.53125,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -10.0625,
+      "step": 1558
+    },
+    {
+      "epoch": 1.9128834355828221,
+      "grad_norm": 16.102556892694622,
+      "learning_rate": 2.9289256198347105e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0634,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.30859375,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -9.0,
+      "step": 1559
+    },
+    {
+      "epoch": 1.9141104294478528,
+      "grad_norm": 8.41879847837055,
+      "learning_rate": 2.9256198347107437e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0306,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.59765625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -10.125,
+      "step": 1560
+    },
+    {
+      "epoch": 1.9153374233128835,
+      "grad_norm": 11.715145668756673,
+      "learning_rate": 2.922314049586777e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -452.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0351,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -10.0625,
+      "step": 1561
+    },
+    {
+      "epoch": 1.9165644171779141,
+      "grad_norm": 6.97755511576966,
+      "learning_rate": 2.9190082644628095e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -436.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0183,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -9.3125,
+      "step": 1562
+    },
+    {
+      "epoch": 1.9177914110429448,
+      "grad_norm": 7.472007517000815,
+      "learning_rate": 2.9157024793388427e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0251,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -9.625,
+      "step": 1563
+    },
+    {
+      "epoch": 1.9190184049079755,
+      "grad_norm": 8.89094542402821,
+      "learning_rate": 2.912396694214876e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0315,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.28125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.1875,
+      "step": 1564
+    },
+    {
+      "epoch": 1.9202453987730062,
+      "grad_norm": 5.401693481934314,
+      "learning_rate": 2.909090909090909e-07,
+      "logits/chosen": -0.39453125,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -322.0,
+      "logps/rejected": -322.0,
+      "loss": 0.0252,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.5078125,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.5625,
+      "step": 1565
+    },
+    {
+      "epoch": 1.9214723926380368,
+      "grad_norm": 17.54721925631407,
+      "learning_rate": 2.905785123966942e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -9.5625,
+      "step": 1566
+    },
+    {
+      "epoch": 1.9226993865030675,
+      "grad_norm": 8.619832489981432,
+      "learning_rate": 2.902479338842975e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0383,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.408203125,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.75,
+      "step": 1567
+    },
+    {
+      "epoch": 1.9239263803680982,
+      "grad_norm": 9.440242433724526,
+      "learning_rate": 2.899173553719008e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -384.0,
+      "loss": 0.028,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -9.25,
+      "step": 1568
+    },
+    {
+      "epoch": 1.9251533742331288,
+      "grad_norm": 8.876395696240671,
+      "learning_rate": 2.895867768595041e-07,
+      "logits/chosen": -0.099609375,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0259,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.408203125,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.3125,
+      "step": 1569
+    },
+    {
+      "epoch": 1.9263803680981595,
+      "grad_norm": 7.999356547168104,
+      "learning_rate": 2.8925619834710744e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0328,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1796875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.25,
+      "step": 1570
+    },
+    {
+      "epoch": 1.9276073619631902,
+      "grad_norm": 11.616155079550802,
+      "learning_rate": 2.889256198347107e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0441,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.1875,
+      "step": 1571
+    },
+    {
+      "epoch": 1.9288343558282208,
+      "grad_norm": 5.433830123431735,
+      "learning_rate": 2.88595041322314e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.5,
+      "logps/chosen": -412.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0176,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.419921875,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -10.1875,
+      "step": 1572
+    },
+    {
+      "epoch": 1.9300613496932515,
+      "grad_norm": 12.21034381437247,
+      "learning_rate": 2.8826446280991734e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0545,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.1826171875,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -9.4375,
+      "step": 1573
+    },
+    {
+      "epoch": 1.9312883435582822,
+      "grad_norm": 13.499082687285052,
+      "learning_rate": 2.8793388429752066e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.40234375,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.125,
+      "step": 1574
+    },
+    {
+      "epoch": 1.9325153374233128,
+      "grad_norm": 15.391849130162866,
+      "learning_rate": 2.87603305785124e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -362.0,
+      "loss": 0.062,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.67578125,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.6875,
+      "step": 1575
+    },
+    {
+      "epoch": 1.9337423312883435,
+      "grad_norm": 8.29560522478457,
+      "learning_rate": 2.8727272727272724e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -364.0,
+      "loss": 0.026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.52734375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -10.25,
+      "step": 1576
+    },
+    {
+      "epoch": 1.9349693251533742,
+      "grad_norm": 10.72329395494961,
+      "learning_rate": 2.8694214876033056e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.98046875,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.5,
+      "step": 1577
+    },
+    {
+      "epoch": 1.9361963190184048,
+      "grad_norm": 16.045872486554956,
+      "learning_rate": 2.866115702479339e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0492,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.0257568359375,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -10.0625,
+      "step": 1578
+    },
+    {
+      "epoch": 1.9374233128834355,
+      "grad_norm": 8.551245323989875,
+      "learning_rate": 2.862809917355372e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0344,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.279296875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.75,
+      "step": 1579
+    },
+    {
+      "epoch": 1.9386503067484662,
+      "grad_norm": 6.503379535878091,
+      "learning_rate": 2.8595041322314046e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -444.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0201,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.828125,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.0625,
+      "step": 1580
+    },
+    {
+      "epoch": 1.9398773006134968,
+      "grad_norm": 8.183695473874987,
+      "learning_rate": 2.856198347107438e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0279,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.75,
+      "step": 1581
+    },
+    {
+      "epoch": 1.9411042944785275,
+      "grad_norm": 11.159282394202965,
+      "learning_rate": 2.852892561983471e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -330.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0554,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.1318359375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -9.125,
+      "step": 1582
+    },
+    {
+      "epoch": 1.9423312883435582,
+      "grad_norm": 15.624281443403351,
+      "learning_rate": 2.849586776859504e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0559,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.92578125,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -9.6875,
+      "step": 1583
+    },
+    {
+      "epoch": 1.9435582822085888,
+      "grad_norm": 14.501392157008612,
+      "learning_rate": 2.8462809917355373e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -436.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0639,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.25,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.1875,
+      "step": 1584
+    },
+    {
+      "epoch": 1.9447852760736195,
+      "grad_norm": 12.361255357871427,
+      "learning_rate": 2.84297520661157e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0276,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.97265625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -10.625,
+      "step": 1585
+    },
+    {
+      "epoch": 1.9460122699386502,
+      "grad_norm": 12.44660865452298,
+      "learning_rate": 2.839669421487603e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0482,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.3125,
+      "step": 1586
+    },
+    {
+      "epoch": 1.9472392638036808,
+      "grad_norm": 10.155482386142491,
+      "learning_rate": 2.8363636363636363e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0395,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.06103515625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.4375,
+      "step": 1587
+    },
+    {
+      "epoch": 1.9484662576687115,
+      "grad_norm": 7.436422121571213,
+      "learning_rate": 2.8330578512396695e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0359,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.44921875,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.3125,
+      "step": 1588
+    },
+    {
+      "epoch": 1.9496932515337422,
+      "grad_norm": 8.76431349347538,
+      "learning_rate": 2.829752066115702e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0282,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2470703125,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -10.4375,
+      "step": 1589
+    },
+    {
+      "epoch": 1.9509202453987728,
+      "grad_norm": 7.189610220189287,
+      "learning_rate": 2.8264462809917353e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0227,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.83203125,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -9.9375,
+      "step": 1590
+    },
+    {
+      "epoch": 1.9521472392638037,
+      "grad_norm": 7.800363828306239,
+      "learning_rate": 2.8231404958677685e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0243,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1572265625,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -10.1875,
+      "step": 1591
+    },
+    {
+      "epoch": 1.9533742331288344,
+      "grad_norm": 6.9621963464144265,
+      "learning_rate": 2.8198347107438017e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0319,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.240234375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.9375,
+      "step": 1592
+    },
+    {
+      "epoch": 1.954601226993865,
+      "grad_norm": 9.330460459092231,
+      "learning_rate": 2.816528925619835e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0303,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.005218505859375,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -10.5625,
+      "step": 1593
+    },
+    {
+      "epoch": 1.9558282208588957,
+      "grad_norm": 13.631057553651138,
+      "learning_rate": 2.8132231404958675e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -328.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0628,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.26953125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.6875,
+      "step": 1594
+    },
+    {
+      "epoch": 1.9570552147239264,
+      "grad_norm": 14.520406844691742,
+      "learning_rate": 2.8099173553719007e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0438,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.287109375,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -10.8125,
+      "step": 1595
+    },
+    {
+      "epoch": 1.958282208588957,
+      "grad_norm": 7.279664367583889,
+      "learning_rate": 2.806611570247934e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0278,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.212890625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -10.25,
+      "step": 1596
+    },
+    {
+      "epoch": 1.9595092024539877,
+      "grad_norm": 11.848236709613104,
+      "learning_rate": 2.803305785123967e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0332,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.349609375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.5,
+      "step": 1597
+    },
+    {
+      "epoch": 1.9607361963190184,
+      "grad_norm": 20.200578986204675,
+      "learning_rate": 2.7999999999999997e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0706,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.28515625,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -10.1875,
+      "step": 1598
+    },
+    {
+      "epoch": 1.961963190184049,
+      "grad_norm": 6.9742046512408855,
+      "learning_rate": 2.796694214876033e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0277,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1044921875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -10.125,
+      "step": 1599
+    },
+    {
+      "epoch": 1.9631901840490797,
+      "grad_norm": 12.052751187248274,
+      "learning_rate": 2.793388429752066e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0386,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.2041015625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.875,
+      "step": 1600
+    },
+    {
+      "epoch": 1.9644171779141104,
+      "grad_norm": 10.394378430820012,
+      "learning_rate": 2.790082644628099e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.25,
+      "logps/chosen": -342.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0494,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.41796875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -9.5,
+      "step": 1601
+    },
+    {
+      "epoch": 1.965644171779141,
+      "grad_norm": 11.67542858875569,
+      "learning_rate": 2.7867768595041324e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -438.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0411,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.5,
+      "step": 1602
+    },
+    {
+      "epoch": 1.9668711656441717,
+      "grad_norm": 9.973844708079088,
+      "learning_rate": 2.783471074380165e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -468.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0408,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.08544921875,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -10.6875,
+      "step": 1603
+    },
+    {
+      "epoch": 1.9680981595092024,
+      "grad_norm": 12.506574586450439,
+      "learning_rate": 2.780165289256198e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0453,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.46484375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.0625,
+      "step": 1604
+    },
+    {
+      "epoch": 1.969325153374233,
+      "grad_norm": 14.4298856829564,
+      "learning_rate": 2.7768595041322314e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -428.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0627,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": -0.171875,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.9375,
+      "step": 1605
+    },
+    {
+      "epoch": 1.9705521472392638,
+      "grad_norm": 10.928070857156225,
+      "learning_rate": 2.7735537190082645e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -414.0,
+      "loss": 0.032,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8984375,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -10.1875,
+      "step": 1606
+    },
+    {
+      "epoch": 1.9717791411042946,
+      "grad_norm": 10.048807722409952,
+      "learning_rate": 2.770247933884297e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.470703125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -344.0,
+      "loss": 0.036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.546875,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.875,
+      "step": 1607
+    },
+    {
+      "epoch": 1.9730061349693253,
+      "grad_norm": 9.370791095894255,
+      "learning_rate": 2.7669421487603304e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0296,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.3671875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -10.0,
+      "step": 1608
+    },
+    {
+      "epoch": 1.974233128834356,
+      "grad_norm": 13.452578592993975,
+      "learning_rate": 2.7636363636363635e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0525,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.263671875,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -10.4375,
+      "step": 1609
+    },
+    {
+      "epoch": 1.9754601226993866,
+      "grad_norm": 17.64298790831311,
+      "learning_rate": 2.7603305785123967e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0446,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.08056640625,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.875,
+      "step": 1610
+    },
+    {
+      "epoch": 1.9766871165644173,
+      "grad_norm": 11.712184783781975,
+      "learning_rate": 2.7570247933884294e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -326.0,
+      "logps/rejected": -322.0,
+      "loss": 0.0566,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.55859375,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -9.3125,
+      "step": 1611
+    },
+    {
+      "epoch": 1.977914110429448,
+      "grad_norm": 11.429670120969462,
+      "learning_rate": 2.7537190082644626e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -440.0,
+      "loss": 0.0371,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.6875,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -10.125,
+      "step": 1612
+    },
+    {
+      "epoch": 1.9791411042944786,
+      "grad_norm": 9.475634561082307,
+      "learning_rate": 2.7504132231404957e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -440.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0259,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.380859375,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -10.9375,
+      "step": 1613
+    },
+    {
+      "epoch": 1.9803680981595093,
+      "grad_norm": 15.195294834798235,
+      "learning_rate": 2.747107438016529e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0612,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.287109375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.5625,
+      "step": 1614
+    },
+    {
+      "epoch": 1.98159509202454,
+      "grad_norm": 17.650174474381533,
+      "learning_rate": 2.743801652892562e-07,
+      "logits/chosen": -0.375,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0668,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.25390625,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -10.6875,
+      "step": 1615
+    },
+    {
+      "epoch": 1.9828220858895707,
+      "grad_norm": 15.324230092096657,
+      "learning_rate": 2.7404958677685947e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0512,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.4609375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -10.0,
+      "step": 1616
+    },
+    {
+      "epoch": 1.9840490797546013,
+      "grad_norm": 19.69575258656083,
+      "learning_rate": 2.737190082644628e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -356.0,
+      "loss": 0.06,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.36328125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -9.3125,
+      "step": 1617
+    },
+    {
+      "epoch": 1.985276073619632,
+      "grad_norm": 8.216862123400189,
+      "learning_rate": 2.733884297520661e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0363,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.0419921875,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.875,
+      "step": 1618
+    },
+    {
+      "epoch": 1.9865030674846627,
+      "grad_norm": 10.46369822241489,
+      "learning_rate": 2.730578512396694e-07,
+      "logits/chosen": -0.330078125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0435,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.20703125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -10.25,
+      "step": 1619
+    },
+    {
+      "epoch": 1.9877300613496933,
+      "grad_norm": 12.60171916853773,
+      "learning_rate": 2.727272727272727e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -362.0,
+      "loss": 0.033,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.00213623046875,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -10.3125,
+      "step": 1620
+    },
+    {
+      "epoch": 1.988957055214724,
+      "grad_norm": 8.365282012776243,
+      "learning_rate": 2.72396694214876e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0317,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.3359375,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -10.25,
+      "step": 1621
+    },
+    {
+      "epoch": 1.9901840490797547,
+      "grad_norm": 11.044384202131916,
+      "learning_rate": 2.720661157024793e-07,
+      "logits/chosen": -0.455078125,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0438,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.1552734375,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -10.5,
+      "step": 1622
+    },
+    {
+      "epoch": 1.9914110429447853,
+      "grad_norm": 8.905193463781423,
+      "learning_rate": 2.7173553719008264e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0235,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.36328125,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -10.375,
+      "step": 1623
+    },
+    {
+      "epoch": 1.992638036809816,
+      "grad_norm": 6.811861007896774,
+      "learning_rate": 2.7140495867768596e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0255,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -10.375,
+      "step": 1624
+    },
+    {
+      "epoch": 1.9938650306748467,
+      "grad_norm": 10.024309031384108,
+      "learning_rate": 2.710743801652892e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -442.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0316,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -10.0,
+      "step": 1625
+    },
+    {
+      "epoch": 1.9950920245398773,
+      "grad_norm": 18.695693680609278,
+      "learning_rate": 2.7074380165289254e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0647,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1953125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.75,
+      "step": 1626
+    },
+    {
+      "epoch": 1.996319018404908,
+      "grad_norm": 10.936992214829806,
+      "learning_rate": 2.7041322314049586e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0422,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.58984375,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -10.0,
+      "step": 1627
+    },
+    {
+      "epoch": 1.9975460122699387,
+      "grad_norm": 9.584409085330941,
+      "learning_rate": 2.700826446280992e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0422,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.1875,
+      "step": 1628
+    },
+    {
+      "epoch": 1.9987730061349693,
+      "grad_norm": 13.453298121808412,
+      "learning_rate": 2.6975206611570244e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0514,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.2060546875,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.8125,
+      "step": 1629
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 11.391241637712861,
+      "learning_rate": 2.6942148760330576e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0366,
+      "rewards/accuracies": 0.9915966987609863,
+      "rewards/chosen": 0.4140625,
+      "rewards/margins": 11.125,
+      "rewards/rejected": -10.6875,
+      "step": 1630
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2445,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-1630/training_args.bin b/checkpoint-1630/training_args.bin
new file mode 100644
index 0000000..70c7ec2
--- /dev/null
+++ b/checkpoint-1630/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:528936595b6d96d00e537c7ae1257bb1454c1a31a8f47041035d37b1fc6d35e1
+size 9528
diff --git a/checkpoint-1630/zero_to_fp32.py b/checkpoint-1630/zero_to_fp32.py
new file mode 100644
index 0000000..24cc342
--- /dev/null
+++ b/checkpoint-1630/zero_to_fp32.py
@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoint-2445/config.json b/checkpoint-2445/config.json
new file mode 100644
index 0000000..ded5a8b
--- /dev/null
+++ b/checkpoint-2445/config.json
@@ -0,0 +1,35 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128009,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 8.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.50.0",
+  "use_cache": false,
+  "vocab_size": 128256
+}
diff --git a/checkpoint-2445/generation_config.json b/checkpoint-2445/generation_config.json
new file mode 100644
index 0000000..0acdb5a
--- /dev/null
+++ b/checkpoint-2445/generation_config.json
@@ -0,0 +1,12 @@
+{
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.50.0"
+}
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..a957fc1
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:495b64ae222376985389064ae6f6b06637173b40c94ed875daf1664deae96fb4
+size 1514015111
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..aafd941
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c48f900c9f1a5b0fea1cdc963a5188912f27efced260fdf803d2adf6e45b1e0f
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..107a2e0
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d14804c211f485c36f77484d777a430c3b914c9303d7598836e73603c83f8c5b
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..10d6986
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46b64c55ae2fc10c4467c498428ca4afe658c1cedb68886199b0b812d275a3d2
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..b7c10f6
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2797d8bb4c99d8a8170c51efa27de192ed8e291be1a87a12db4f7fdd2f0c790b
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..713d662
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96797925ddfa0ea84ac16ed0173c0c9b2846ae8652528912ce7c01bd66b4eabf
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..084e81a
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17653d591b10611c3a48f9898681b89d76acce5196378deb99be2f6fc82dc802
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..437e2a7
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29919fd74e9c00732b05652e3d468a9b4b12e75b2982953ff9421e86d44ba410
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..8c4dfa5
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40a1f9b7c30f57e6b4b5419eb921940da7371bc7489c018ad69d2399357d3fd3
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..5e934c1
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83da36339302cf9622dbfd31d150306001dc4b34b06969d04a7894565f6bc283
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..6d4f0e2
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14f78f053ecfb35f93e9a2cd96435c3766d3aa6003d97fd8304e89d601255154
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..cd5b5ae
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c135a23aa558c1691e470bd51a863e81ecdf2eba55374dfcad97b0d206009ea6
+size 1514015111
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..fcb6d44
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad00b2188240966d60649e54cb31cab58aff00444aa9d4ae71734d26421e9522
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..31bd834
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19c8b2c5ec9dec7c225444d2c24ff8d45fbab0f8dc27b413f733c7f95ac0ca90
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..8bf2f94
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f87432df5d6abf2ee643941560d13e13971e35b00adbd47230ff83695e04bada
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..3ced712
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5149802f91eb49cd70b715d82f1ead4818043a739e3fdd6fc632b54e0a68cd7
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..c550e03
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:74db9146e3af5c2ebe40f9d6231796c3e87726971b6f90d3153a909a23b4e30d
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..d355c4f
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94f7b79fef776a4a020982048223cdc8c0b36451e740a48e4bdcc04a33be9bf4
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..0b92904
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c453d87bc26655c21be52200bd9bd74b6b71cf7172e74dc67814ac877bf3013
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..1007271
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07cbffb47ac817fd1da3ef0ff336f185e32156f227633704a4a92a2eee7b650c
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..3b6182c
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dcdd1c877bdb007a298f96750d28124df2ddbf4a6dee76afb1f27023e1555768
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..642e1a4
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34969048c56cf1f41988dd4e262a28414cff85c9b24539405899fb78e2cc1bcc
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..2533dd2
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:900395bb0c73a7d72157056a57cb776914321ad930764a44736d571654b76aaf
+size 1514015111
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..f5272b6
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67c6d6c3dcd2354edde1390a61f3536e84739af12933dca3079872beaabe8b34
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..3328c24
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5807e6a8f5c4ae8ce895422f16847f9d6440772efa66a835a42f1adfd57a5029
+size 1514016442
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..e35f49b
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38f9417b1e966cbeb5ae7cfe1176bc63bd4415cdfac0f90055548ed97bf4e69f
+size 1514015111
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..256430a
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9be923b473f71e01d20546bf68436e05627238c56234a1eac5d38cfb5b4c4f53
+size 1514015111
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..b6ec845
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1bf6b5eb7e325e0003fe2837d6237bed34b0760a6ec99a23200ee8e83246d5c
+size 1514015111
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..2edc23e
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2fbe32130c2999aea0d8c22861891be926cdfbb73d2f293d5769e636a3ed5884
+size 1514015111
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..fc8d34f
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7178ddd81b5c0ee0a1e0173473de0429da23e0e72c278acc1acc12e4709db6a
+size 1514015111
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..cf49f51
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:79a3e125a80f9c1261b62944cd57b612d97abc91ad1d31f9ed4439f9816ffe35
+size 1514015111
diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..d6c379e
--- /dev/null
+++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:458bb1945f4ebfdc8a7f65a389f4530be514113b740ac0eef27e9bd72a97974c
+size 1514015111
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_0_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_0_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..b74ff49
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_0_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ffe2ec38edbe0b5d6e6150e6e9c4874bad566b9a275825fae36b115dfb705507
+size 150245
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_10_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_10_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..9f111a7
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_10_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4e658b292c123f55b3d55c45e964bd863e856250b17842c831ffef0c85f16a1
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_11_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_11_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..06fe07b
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_11_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9fecdf063c4da0cf60d7f6b35ac9ea6a055abe701f722bc2c3d2bb7b044c2ea1
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_12_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_12_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..a7b1e94
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_12_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bfc82216501ff83ea47d43f9f20b16bc0d56eb698d65763fa53af58261283b40
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_13_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_13_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..76a6636
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_13_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5ca14f3ce2aa6e8432025d9bcfae023f4028772903aedf27251a73ee89d9d9c
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_14_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_14_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..1a1c1d9
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_14_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b499c7332ae7a32106b346ae639177132aedc19afabb2738ca0e64bc3d7de2c1
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_15_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_15_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..cc19f64
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_15_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:329e41b68c60e83e871a716f6082a4be3c906e00da7ae73e051bd4337426c778
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_16_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_16_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..413b1e6
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_16_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2100aeaab7b7a3f79e3a5f65f4b0921ccb135f81fa1e051446156c3a3bfd73bb
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_17_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_17_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..0f2822c
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_17_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2fbc8d92199cf5ea2edfdebf931db8adfe2a5f8765c18e9d8179778be59c539
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_18_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_18_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..7601315
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_18_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3db66e4fe950e5c4b5246e87074224980c95347d1ba107cbfc86679f6d97445d
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_19_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_19_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..85f488c
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_19_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8fb63cd572f258454684e6ca311b00501e7d86139f94138968c00257036fa155
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_1_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_1_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..4ae6595
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_1_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bae542fcea121575e03aef13180259386db45d8338cfb8e170c1b1e936f07f56
+size 150245
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_20_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_20_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..7af3882
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_20_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5868fd1a75aa1dd0d95ee8c620b62c2278b143828cbebe4eccd1407be9677f7b
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_21_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_21_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..40ace16
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_21_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ddc82f8cf5a751cf0e3262640ed49766df68af472a6d4814e51dac88862d27d
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_22_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_22_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..a4ae388
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_22_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:12f2337e40e00274ed590c1fe24b40149ea7e94cffbef40a7cbf47645389cb3f
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_23_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_23_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..26d41ca
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_23_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d4886a3dfdec0f0415d38692a739a753ab7e29dc88fa108e8b6fe42625ef192
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_24_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_24_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..1e1c670
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_24_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5fb4ecdb0b270ff0eced437c781d31b0cb014809d32917e168940877bfb03f4
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_25_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_25_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..f2774b4
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_25_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0901e901fe596b3138d2ccc32204539f794ef6f279c6b42be27a40c895c787d3
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_26_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_26_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..7e7363b
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_26_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f207ac4a816c43bf844c2c8c80644cbc46e16a4111a38c58827903cf3d4bcdd
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_27_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_27_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..cbdd845
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_27_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd23381e14373662f70a9dce1c0eeb15924cb8077c967ddc0e6d1e1a572ce43c
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_28_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_28_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..b4a85c8
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_28_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:059bf1ba37ee92ccca47187fded153a9af22ee7044a0cee33b17a8c25c31b2c1
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_29_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_29_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..387d79d
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_29_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3f05ba702f76a781cc0cca9f4d3fab13ea3f597fc2ea91a5201fba24d0419b0
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_2_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_2_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..d0b24ac
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_2_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0bf6f1b5fd7f395e28561eb3210bc8f3fa0d6cb589439ed820880d6b6a48bd6b
+size 150245
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_30_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_30_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..0f81b1e
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_30_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60601369222e025ec7ace86e4faa98f72f42033dbddedc18dcc9e4a05d45b2c3
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_31_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_31_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..94614c4
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_31_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6cf71cecae2998dfe48359806530fde46071708a330bf5da58ac545c673467de
+size 150540
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_3_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_3_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..ad35869
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_3_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e70e74735403d741bd55d1ece22450b62be54ef37652769657aa404eed47d2ee
+size 150245
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_4_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_4_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..64d925c
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_4_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4991f0ead2f78d5715727716ecf7bfe732ecfd7ab68f34f25f15fcc68c5a6267
+size 150245
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_5_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_5_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..cbcf333
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_5_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5497e5c0a1f39abf9e032be83d4142aaca609dcce04ee031ffb3d11fd1e3c1a
+size 150245
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_6_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_6_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..fdd724e
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_6_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b97d0b7bc65b7c10b9af30baaecc9e58d79b2cb0861ae27187b4b0b3ba2674b
+size 150245
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_7_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_7_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..2b06dbe
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_7_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e2de5673e2d582e3de81f7b93016335f067b70ac5717d7651cb85afa1756486
+size 150245
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_8_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_8_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..83d25ae
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_8_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d89c67316075467476c1eb25aa00637bc97f5c0b241d464e6ad18d5544dc314b
+size 150245
diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_9_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_9_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..ee1482a
--- /dev/null
+++ b/checkpoint-2445/global_step2445/zero_pp_rank_9_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:909fce6df9e30ff0e3b1913a1313d8414cff3352514e4d46eb424a280c594a27
+size 150245
diff --git a/checkpoint-2445/latest b/checkpoint-2445/latest
new file mode 100644
index 0000000..5247c35
--- /dev/null
+++ b/checkpoint-2445/latest
@@ -0,0 +1 @@
+global_step2445
\ No newline at end of file
diff --git a/checkpoint-2445/model-00001-of-00004.safetensors b/checkpoint-2445/model-00001-of-00004.safetensors
new file mode 100644
index 0000000..8672400
--- /dev/null
+++ b/checkpoint-2445/model-00001-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45f844c7bcec934a51bb941e54d30a655fc8db45dd739d36738f2fa4ce1de909
+size 4976698672
diff --git a/checkpoint-2445/model-00002-of-00004.safetensors b/checkpoint-2445/model-00002-of-00004.safetensors
new file mode 100644
index 0000000..5ef40e7
--- /dev/null
+++ b/checkpoint-2445/model-00002-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e194d3a55358b5c5d86589c20ee9b494530bf727cc1955b7907610cd4b044682
+size 4999802720
diff --git a/checkpoint-2445/model-00003-of-00004.safetensors b/checkpoint-2445/model-00003-of-00004.safetensors
new file mode 100644
index 0000000..2715e04
--- /dev/null
+++ b/checkpoint-2445/model-00003-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:774b89d412ab7fd49b257aa6ea2b15450e5a7caaff911791ce76707c059730b3
+size 4915916176
diff --git a/checkpoint-2445/model-00004-of-00004.safetensors b/checkpoint-2445/model-00004-of-00004.safetensors
new file mode 100644
index 0000000..f3b4e77
--- /dev/null
+++ b/checkpoint-2445/model-00004-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:933eca7410a1d89ab7601cb2b6a52d50488ac67c7c5bd30235b9802488e85d19
+size 1168138808
diff --git a/checkpoint-2445/model.safetensors.index.json b/checkpoint-2445/model.safetensors.index.json
new file mode 100644
index 0000000..0fd8120
--- /dev/null
+++ b/checkpoint-2445/model.safetensors.index.json
@@ -0,0 +1,298 @@
+{
+  "metadata": {
+    "total_size": 16060522496
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.norm.weight": "model-00004-of-00004.safetensors"
+  }
+}
diff --git a/checkpoint-2445/rng_state_0.pth b/checkpoint-2445/rng_state_0.pth
new file mode 100644
index 0000000..3525225
--- /dev/null
+++ b/checkpoint-2445/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:08282b46825aa78d10fe10e3fea89555c5b5a691b261a3ddfd58fcb58370edff
+size 15984
diff --git a/checkpoint-2445/rng_state_1.pth b/checkpoint-2445/rng_state_1.pth
new file mode 100644
index 0000000..eeb7d8d
--- /dev/null
+++ b/checkpoint-2445/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbab71d98a3a9a92df82a6bba463947327c3a1bcf35cd9f4f46114641fc42dd9
+size 15984
diff --git a/checkpoint-2445/rng_state_10.pth b/checkpoint-2445/rng_state_10.pth
new file mode 100644
index 0000000..338c1db
--- /dev/null
+++ b/checkpoint-2445/rng_state_10.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54e5c2d65c14df39137f6a3bd8314f534dad93d72efe8e40e23a61187ba74ce8
+size 15997
diff --git a/checkpoint-2445/rng_state_11.pth b/checkpoint-2445/rng_state_11.pth
new file mode 100644
index 0000000..4bbb9bf
--- /dev/null
+++ b/checkpoint-2445/rng_state_11.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f578146bcaaed333cf4637aa3b454e2f6238e691e45b4160001b63b4f8b5ce4e
+size 15997
diff --git a/checkpoint-2445/rng_state_12.pth b/checkpoint-2445/rng_state_12.pth
new file mode 100644
index 0000000..82f488f
--- /dev/null
+++ b/checkpoint-2445/rng_state_12.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8463662162a837871bb50af43a407175ca95f3f364bc13c7d065e8b6929bad11
+size 15997
diff --git a/checkpoint-2445/rng_state_13.pth b/checkpoint-2445/rng_state_13.pth
new file mode 100644
index 0000000..1b4dd63
--- /dev/null
+++ b/checkpoint-2445/rng_state_13.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe9bafd75244b907f13aa5b102963f4d018d7b80b94ddb0262313774362d9305
+size 15997
diff --git a/checkpoint-2445/rng_state_14.pth b/checkpoint-2445/rng_state_14.pth
new file mode 100644
index 0000000..b7f3c97
--- /dev/null
+++ b/checkpoint-2445/rng_state_14.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5e353c1ca54c464e3464d695507da7da6c134e4ad9d6acab6f72e52d5bf13c7
+size 15997
diff --git a/checkpoint-2445/rng_state_15.pth b/checkpoint-2445/rng_state_15.pth
new file mode 100644
index 0000000..8e4dee4
--- /dev/null
+++ b/checkpoint-2445/rng_state_15.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24f1ef20cfc904ac8f811ceffab91f00cccadca0e878c8ef8a7335e83ef9513f
+size 15997
diff --git a/checkpoint-2445/rng_state_16.pth b/checkpoint-2445/rng_state_16.pth
new file mode 100644
index 0000000..4033ee0
--- /dev/null
+++ b/checkpoint-2445/rng_state_16.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cecf995eb1b13a85a673384ed67c75acc989e9d373ac7da2d90cbc7af2c1b4a7
+size 15997
diff --git a/checkpoint-2445/rng_state_17.pth b/checkpoint-2445/rng_state_17.pth
new file mode 100644
index 0000000..0a50e02
--- /dev/null
+++ b/checkpoint-2445/rng_state_17.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:375986c2a348f319aeed3322fcf23be7787f801c585046ff217d7855463fbc0d
+size 15997
diff --git a/checkpoint-2445/rng_state_18.pth b/checkpoint-2445/rng_state_18.pth
new file mode 100644
index 0000000..075e295
--- /dev/null
+++ b/checkpoint-2445/rng_state_18.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a50dfdcb902d78e55850cabc410552a4e651090792c2e086926cbda383b4f35f
+size 15997
diff --git a/checkpoint-2445/rng_state_19.pth b/checkpoint-2445/rng_state_19.pth
new file mode 100644
index 0000000..ac69e75
--- /dev/null
+++ b/checkpoint-2445/rng_state_19.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:79850d41cfed8d83ef96222b2b956a0ec273b7fa78630d3cb0e86f2ab29bb934
+size 15997
diff --git a/checkpoint-2445/rng_state_2.pth b/checkpoint-2445/rng_state_2.pth
new file mode 100644
index 0000000..e144a44
--- /dev/null
+++ b/checkpoint-2445/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:caac82d57d878d30219a4f9ec289a97ff90c53afc160b968f251b3fd3454b8d8
+size 15984
diff --git a/checkpoint-2445/rng_state_20.pth b/checkpoint-2445/rng_state_20.pth
new file mode 100644
index 0000000..f4588d3
--- /dev/null
+++ b/checkpoint-2445/rng_state_20.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60201c3050f5fddd8ba67c9cc8ea589b8296872acef041a46972a75489ba48f2
+size 15997
diff --git a/checkpoint-2445/rng_state_21.pth b/checkpoint-2445/rng_state_21.pth
new file mode 100644
index 0000000..94a9f24
--- /dev/null
+++ b/checkpoint-2445/rng_state_21.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b45ccf4c1d4fa82ab73d5b1084c8895bb7a843bb65c39bb0b2bb2a6192e6df3
+size 15997
diff --git a/checkpoint-2445/rng_state_22.pth b/checkpoint-2445/rng_state_22.pth
new file mode 100644
index 0000000..7d803e4
--- /dev/null
+++ b/checkpoint-2445/rng_state_22.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54521e12e4ad123a5cdba9d3c092e63ee37b099ff802bbe417dfb47861244ab9
+size 15997
diff --git a/checkpoint-2445/rng_state_23.pth b/checkpoint-2445/rng_state_23.pth
new file mode 100644
index 0000000..10c1a68
--- /dev/null
+++ b/checkpoint-2445/rng_state_23.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4917fcd537e822a9012f654e3131d5dbd358e1916dec368f464d89c1223a2e5
+size 15997
diff --git a/checkpoint-2445/rng_state_24.pth b/checkpoint-2445/rng_state_24.pth
new file mode 100644
index 0000000..67c352c
--- /dev/null
+++ b/checkpoint-2445/rng_state_24.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6b078914d8fc5b3568727c1679a97889d10f35460aaa0147c1ab343e0a8576f
+size 15997
diff --git a/checkpoint-2445/rng_state_25.pth b/checkpoint-2445/rng_state_25.pth
new file mode 100644
index 0000000..4e42fe3
--- /dev/null
+++ b/checkpoint-2445/rng_state_25.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58235624c008372a4489d1dc4d4a5d582e83b5954529d532a8ac9e27cc1117eb
+size 15997
diff --git a/checkpoint-2445/rng_state_26.pth b/checkpoint-2445/rng_state_26.pth
new file mode 100644
index 0000000..58fc36f
--- /dev/null
+++ b/checkpoint-2445/rng_state_26.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6d0a9e394ce0c35df215df576763794e20589536fd7d2e8aae0d2bfaac14eb95
+size 15997
diff --git a/checkpoint-2445/rng_state_27.pth b/checkpoint-2445/rng_state_27.pth
new file mode 100644
index 0000000..0e9b2ee
--- /dev/null
+++ b/checkpoint-2445/rng_state_27.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a40fc515e550756e1b3497c9e62fa179918e76adb87ab3233db714b992268c8
+size 15997
diff --git a/checkpoint-2445/rng_state_28.pth b/checkpoint-2445/rng_state_28.pth
new file mode 100644
index 0000000..befeed4
--- /dev/null
+++ b/checkpoint-2445/rng_state_28.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6dadf1f4ae507c5e5ab940441273a3552a7e59fcc516e7dfddd1956d1c5c5a3d
+size 15997
diff --git a/checkpoint-2445/rng_state_29.pth b/checkpoint-2445/rng_state_29.pth
new file mode 100644
index 0000000..dfca012
--- /dev/null
+++ b/checkpoint-2445/rng_state_29.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73dfc967a343045291988f3bcbbdee93e0877bb869db48fd675e9946c2c43507
+size 15997
diff --git a/checkpoint-2445/rng_state_3.pth b/checkpoint-2445/rng_state_3.pth
new file mode 100644
index 0000000..a10f352
--- /dev/null
+++ b/checkpoint-2445/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19762d2d370222b01817da11bbaa6665d542293373186d66f754e7246bb861ed
+size 15984
diff --git a/checkpoint-2445/rng_state_30.pth b/checkpoint-2445/rng_state_30.pth
new file mode 100644
index 0000000..a6f6699
--- /dev/null
+++ b/checkpoint-2445/rng_state_30.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6890d56ed291634fd39a3901c77e10c249a355684a11b3e34af07a400ff9189c
+size 15997
diff --git a/checkpoint-2445/rng_state_31.pth b/checkpoint-2445/rng_state_31.pth
new file mode 100644
index 0000000..5c916cb
--- /dev/null
+++ b/checkpoint-2445/rng_state_31.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25d52ac86e2e15b3ec76429e008562b498d6431769f7f3ad24c182d032f7701b
+size 15997
diff --git a/checkpoint-2445/rng_state_4.pth b/checkpoint-2445/rng_state_4.pth
new file mode 100644
index 0000000..e6ef215
--- /dev/null
+++ b/checkpoint-2445/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00c7508b346a7d3c5c23392845f1d013331114ade778794b76e919cb3ed5d33e
+size 15984
diff --git a/checkpoint-2445/rng_state_5.pth b/checkpoint-2445/rng_state_5.pth
new file mode 100644
index 0000000..126662e
--- /dev/null
+++ b/checkpoint-2445/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b89de7d14dd20a191f56b74c816ef8b7fe5c171e31efbeadbf321c4539ed68c3
+size 15984
diff --git a/checkpoint-2445/rng_state_6.pth b/checkpoint-2445/rng_state_6.pth
new file mode 100644
index 0000000..d4e6b27
--- /dev/null
+++ b/checkpoint-2445/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c71152053553e6e22d670fbc4fd7550bf8a046b54cad7b71869787986a6a42c
+size 15984
diff --git a/checkpoint-2445/rng_state_7.pth b/checkpoint-2445/rng_state_7.pth
new file mode 100644
index 0000000..14e214a
--- /dev/null
+++ b/checkpoint-2445/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b67db12a26a26ffe03d9afc84a43857eb2e5b2fec2dd189653b415f74208190
+size 15984
diff --git a/checkpoint-2445/rng_state_8.pth b/checkpoint-2445/rng_state_8.pth
new file mode 100644
index 0000000..22cf6c8
--- /dev/null
+++ b/checkpoint-2445/rng_state_8.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50d06b29f11da5c2e4a388e6620296a543a68f4d46bb2bb978188a8e374f8925
+size 15984
diff --git a/checkpoint-2445/rng_state_9.pth b/checkpoint-2445/rng_state_9.pth
new file mode 100644
index 0000000..a4fd554
--- /dev/null
+++ b/checkpoint-2445/rng_state_9.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bde660db671106d0d18882a4db0faf4fe72837df15f5537708435e1d684a549e
+size 15984
diff --git a/checkpoint-2445/scheduler.pt b/checkpoint-2445/scheduler.pt
new file mode 100644
index 0000000..42c5cd5
--- /dev/null
+++ b/checkpoint-2445/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ce6f944c736db62a752c1b4cfd0f2e31025a1d0ed2e97a4aa72c7f1104b1c79
+size 1064
diff --git a/checkpoint-2445/special_tokens_map.json b/checkpoint-2445/special_tokens_map.json
new file mode 100644
index 0000000..278b7f0
--- /dev/null
+++ b/checkpoint-2445/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-2445/tokenizer.json b/checkpoint-2445/tokenizer.json
new file mode 100644
index 0000000..1c1d8d5
--- /dev/null
+++ b/checkpoint-2445/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-2445/tokenizer_config.json b/checkpoint-2445/tokenizer_config.json
new file mode 100644
index 0000000..8098161
--- /dev/null
+++ b/checkpoint-2445/tokenizer_config.json
@@ -0,0 +1,2064 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "<|end_of_text|>",
+  "tokenizer_class": "PreTrainedTokenizer"
+}
diff --git a/checkpoint-2445/trainer_state.json b/checkpoint-2445/trainer_state.json
new file mode 100644
index 0000000..efab010
--- /dev/null
+++ b/checkpoint-2445/trainer_state.json
@@ -0,0 +1,36709 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 2445,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001226993865030675,
+      "grad_norm": 151.42932274482283,
+      "learning_rate": 3.2e-08,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -474.0,
+      "logps/rejected": -316.0,
+      "loss": 0.6914,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.00245398773006135,
+      "grad_norm": 162.73344539002164,
+      "learning_rate": 6.4e-08,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -292.0,
+      "loss": 0.6914,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 2
+    },
+    {
+      "epoch": 0.0036809815950920245,
+      "grad_norm": 111.8042794987884,
+      "learning_rate": 9.599999999999999e-08,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -272.0,
+      "loss": 0.6886,
+      "rewards/accuracies": 0.3046875,
+      "rewards/chosen": 0.01531982421875,
+      "rewards/margins": 0.01361083984375,
+      "rewards/rejected": 0.0016632080078125,
+      "step": 3
+    },
+    {
+      "epoch": 0.0049079754601227,
+      "grad_norm": 155.41451533543852,
+      "learning_rate": 1.28e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -242.0,
+      "loss": 0.7131,
+      "rewards/accuracies": 0.328125,
+      "rewards/chosen": -0.00408935546875,
+      "rewards/margins": -0.0218505859375,
+      "rewards/rejected": 0.017822265625,
+      "step": 4
+    },
+    {
+      "epoch": 0.006134969325153374,
+      "grad_norm": 307.019899938903,
+      "learning_rate": 1.6e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -270.0,
+      "loss": 0.6906,
+      "rewards/accuracies": 0.3515625,
+      "rewards/chosen": 0.01171875,
+      "rewards/margins": 0.01251220703125,
+      "rewards/rejected": -0.00078582763671875,
+      "step": 5
+    },
+    {
+      "epoch": 0.007361963190184049,
+      "grad_norm": 340.66722515261114,
+      "learning_rate": 1.9199999999999997e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -306.0,
+      "loss": 0.7151,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.0228271484375,
+      "rewards/margins": -0.034912109375,
+      "rewards/rejected": 0.01220703125,
+      "step": 6
+    },
+    {
+      "epoch": 0.008588957055214725,
+      "grad_norm": 239.24556126143491,
+      "learning_rate": 2.2400000000000002e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -266.0,
+      "loss": 0.7032,
+      "rewards/accuracies": 0.2734375,
+      "rewards/chosen": -0.0003414154052734375,
+      "rewards/margins": -0.01373291015625,
+      "rewards/rejected": 0.01336669921875,
+      "step": 7
+    },
+    {
+      "epoch": 0.0098159509202454,
+      "grad_norm": 139.49365426176607,
+      "learning_rate": 2.56e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -292.0,
+      "loss": 0.7012,
+      "rewards/accuracies": 0.3671875,
+      "rewards/chosen": 0.0284423828125,
+      "rewards/margins": -0.00927734375,
+      "rewards/rejected": 0.037841796875,
+      "step": 8
+    },
+    {
+      "epoch": 0.011042944785276074,
+      "grad_norm": 122.82829055069632,
+      "learning_rate": 2.88e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -294.0,
+      "loss": 0.6871,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.037353515625,
+      "rewards/margins": 0.0244140625,
+      "rewards/rejected": 0.01300048828125,
+      "step": 9
+    },
+    {
+      "epoch": 0.012269938650306749,
+      "grad_norm": 137.4984961934261,
+      "learning_rate": 3.2e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -278.0,
+      "loss": 0.6605,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": 0.057861328125,
+      "rewards/margins": 0.07275390625,
+      "rewards/rejected": -0.01495361328125,
+      "step": 10
+    },
+    {
+      "epoch": 0.013496932515337423,
+      "grad_norm": 154.04829614927058,
+      "learning_rate": 3.52e-07,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -296.0,
+      "loss": 0.6512,
+      "rewards/accuracies": 0.4765625,
+      "rewards/chosen": 0.0625,
+      "rewards/margins": 0.09912109375,
+      "rewards/rejected": -0.036376953125,
+      "step": 11
+    },
+    {
+      "epoch": 0.014723926380368098,
+      "grad_norm": 292.49015304309825,
+      "learning_rate": 3.8399999999999994e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -262.0,
+      "loss": 0.6321,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": 0.080078125,
+      "rewards/margins": 0.1435546875,
+      "rewards/rejected": -0.06298828125,
+      "step": 12
+    },
+    {
+      "epoch": 0.015950920245398775,
+      "grad_norm": 179.53722097804638,
+      "learning_rate": 4.16e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -249.0,
+      "loss": 0.6194,
+      "rewards/accuracies": 0.6171875,
+      "rewards/chosen": 0.12109375,
+      "rewards/margins": 0.1845703125,
+      "rewards/rejected": -0.0634765625,
+      "step": 13
+    },
+    {
+      "epoch": 0.01717791411042945,
+      "grad_norm": 108.27195029021392,
+      "learning_rate": 4.4800000000000004e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -472.0,
+      "logps/rejected": -312.0,
+      "loss": 0.5477,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": 0.310546875,
+      "rewards/margins": 0.408203125,
+      "rewards/rejected": -0.09814453125,
+      "step": 14
+    },
+    {
+      "epoch": 0.018404907975460124,
+      "grad_norm": 71.36850002174661,
+      "learning_rate": 4.8e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.18359375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -290.0,
+      "loss": 0.5893,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": 0.2412109375,
+      "rewards/margins": 0.337890625,
+      "rewards/rejected": -0.09619140625,
+      "step": 15
+    },
+    {
+      "epoch": 0.0196319018404908,
+      "grad_norm": 80.69275734688873,
+      "learning_rate": 5.12e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -306.0,
+      "loss": 0.5357,
+      "rewards/accuracies": 0.6484375,
+      "rewards/chosen": 0.318359375,
+      "rewards/margins": 0.48828125,
+      "rewards/rejected": -0.1689453125,
+      "step": 16
+    },
+    {
+      "epoch": 0.020858895705521473,
+      "grad_norm": 103.00148200369824,
+      "learning_rate": 5.44e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -255.0,
+      "loss": 0.5279,
+      "rewards/accuracies": 0.6953125,
+      "rewards/chosen": 0.369140625,
+      "rewards/margins": 0.54296875,
+      "rewards/rejected": -0.1728515625,
+      "step": 17
+    },
+    {
+      "epoch": 0.022085889570552148,
+      "grad_norm": 90.51865254706846,
+      "learning_rate": 5.76e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -276.0,
+      "loss": 0.5097,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 0.419921875,
+      "rewards/margins": 0.66015625,
+      "rewards/rejected": -0.23828125,
+      "step": 18
+    },
+    {
+      "epoch": 0.023312883435582823,
+      "grad_norm": 63.171857373704086,
+      "learning_rate": 6.079999999999999e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -290.0,
+      "loss": 0.5237,
+      "rewards/accuracies": 0.6953125,
+      "rewards/chosen": 0.4375,
+      "rewards/margins": 0.80078125,
+      "rewards/rejected": -0.36328125,
+      "step": 19
+    },
+    {
+      "epoch": 0.024539877300613498,
+      "grad_norm": 44.73691218363087,
+      "learning_rate": 6.4e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -462.0,
+      "logps/rejected": -314.0,
+      "loss": 0.4345,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.82421875,
+      "rewards/margins": 1.3984375,
+      "rewards/rejected": -0.5703125,
+      "step": 20
+    },
+    {
+      "epoch": 0.025766871165644172,
+      "grad_norm": 60.19744308670884,
+      "learning_rate": 6.72e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -262.0,
+      "loss": 0.4418,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 1.1328125,
+      "rewards/rejected": -0.50390625,
+      "step": 21
+    },
+    {
+      "epoch": 0.026993865030674847,
+      "grad_norm": 60.06635545400663,
+      "learning_rate": 7.04e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -294.0,
+      "loss": 0.4638,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 0.640625,
+      "rewards/margins": 1.25,
+      "rewards/rejected": -0.61328125,
+      "step": 22
+    },
+    {
+      "epoch": 0.02822085889570552,
+      "grad_norm": 81.66202093067754,
+      "learning_rate": 7.36e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -306.0,
+      "loss": 0.5093,
+      "rewards/accuracies": 0.7421875,
+      "rewards/chosen": 0.81640625,
+      "rewards/margins": 1.421875,
+      "rewards/rejected": -0.6015625,
+      "step": 23
+    },
+    {
+      "epoch": 0.029447852760736196,
+      "grad_norm": 83.23689816485307,
+      "learning_rate": 7.679999999999999e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -278.0,
+      "loss": 0.511,
+      "rewards/accuracies": 0.7265625,
+      "rewards/chosen": 0.47265625,
+      "rewards/margins": 1.03125,
+      "rewards/rejected": -0.5546875,
+      "step": 24
+    },
+    {
+      "epoch": 0.03067484662576687,
+      "grad_norm": 38.90986069898983,
+      "learning_rate": 8e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -434.0,
+      "logps/rejected": -300.0,
+      "loss": 0.4067,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.8515625,
+      "rewards/margins": 1.7578125,
+      "rewards/rejected": -0.90625,
+      "step": 25
+    },
+    {
+      "epoch": 0.03190184049079755,
+      "grad_norm": 47.94846803093583,
+      "learning_rate": 7.996694214876033e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -276.0,
+      "loss": 0.4388,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 1.84375,
+      "rewards/rejected": -0.8203125,
+      "step": 26
+    },
+    {
+      "epoch": 0.033128834355828224,
+      "grad_norm": 44.30203970039667,
+      "learning_rate": 7.993388429752066e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -280.0,
+      "loss": 0.3991,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 2.234375,
+      "rewards/rejected": -1.0,
+      "step": 27
+    },
+    {
+      "epoch": 0.0343558282208589,
+      "grad_norm": 38.710283535206564,
+      "learning_rate": 7.990082644628099e-07,
+      "logits/chosen": -0.06591796875,
+      "logits/rejected": -0.1767578125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -262.0,
+      "loss": 0.4113,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 2.015625,
+      "rewards/rejected": -0.74609375,
+      "step": 28
+    },
+    {
+      "epoch": 0.03558282208588957,
+      "grad_norm": 52.910805586215126,
+      "learning_rate": 7.986776859504132e-07,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -322.0,
+      "loss": 0.392,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 2.546875,
+      "rewards/rejected": -1.1015625,
+      "step": 29
+    },
+    {
+      "epoch": 0.03680981595092025,
+      "grad_norm": 45.21313329781316,
+      "learning_rate": 7.983471074380164e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -334.0,
+      "logps/rejected": -266.0,
+      "loss": 0.4322,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 2.03125,
+      "rewards/rejected": -0.93359375,
+      "step": 30
+    },
+    {
+      "epoch": 0.03803680981595092,
+      "grad_norm": 41.258011933772664,
+      "learning_rate": 7.980165289256198e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -276.0,
+      "loss": 0.4115,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 2.765625,
+      "rewards/rejected": -1.1171875,
+      "step": 31
+    },
+    {
+      "epoch": 0.0392638036809816,
+      "grad_norm": 39.16160748489701,
+      "learning_rate": 7.976859504132231e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3836,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5859375,
+      "rewards/margins": 2.71875,
+      "rewards/rejected": -1.1328125,
+      "step": 32
+    },
+    {
+      "epoch": 0.04049079754601227,
+      "grad_norm": 38.84939653302124,
+      "learning_rate": 7.973553719008264e-07,
+      "logits/chosen": -0.0869140625,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -328.0,
+      "logps/rejected": -234.0,
+      "loss": 0.3966,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 2.234375,
+      "rewards/rejected": -1.0859375,
+      "step": 33
+    },
+    {
+      "epoch": 0.04171779141104295,
+      "grad_norm": 44.221302785049176,
+      "learning_rate": 7.970247933884297e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -302.0,
+      "loss": 0.4038,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 2.859375,
+      "rewards/rejected": -1.46875,
+      "step": 34
+    },
+    {
+      "epoch": 0.04294478527607362,
+      "grad_norm": 46.59951208776938,
+      "learning_rate": 7.96694214876033e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -264.0,
+      "loss": 0.3758,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 2.71875,
+      "rewards/rejected": -1.6484375,
+      "step": 35
+    },
+    {
+      "epoch": 0.044171779141104296,
+      "grad_norm": 41.317139252172254,
+      "learning_rate": 7.963636363636364e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3557,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 3.234375,
+      "rewards/rejected": -1.7421875,
+      "step": 36
+    },
+    {
+      "epoch": 0.04539877300613497,
+      "grad_norm": 39.795000825479654,
+      "learning_rate": 7.960330578512397e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3698,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 3.328125,
+      "rewards/rejected": -1.53125,
+      "step": 37
+    },
+    {
+      "epoch": 0.046625766871165646,
+      "grad_norm": 35.75368356475873,
+      "learning_rate": 7.95702479338843e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3582,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 3.6875,
+      "rewards/rejected": -1.671875,
+      "step": 38
+    },
+    {
+      "epoch": 0.04785276073619632,
+      "grad_norm": 46.53573926538421,
+      "learning_rate": 7.953719008264462e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3835,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 3.671875,
+      "rewards/rejected": -1.984375,
+      "step": 39
+    },
+    {
+      "epoch": 0.049079754601226995,
+      "grad_norm": 33.56284839867345,
+      "learning_rate": 7.950413223140495e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3042,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 3.765625,
+      "rewards/rejected": -2.109375,
+      "step": 40
+    },
+    {
+      "epoch": 0.05030674846625767,
+      "grad_norm": 39.46235435607716,
+      "learning_rate": 7.947107438016528e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -278.0,
+      "loss": 0.3708,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 3.546875,
+      "rewards/rejected": -2.0,
+      "step": 41
+    },
+    {
+      "epoch": 0.051533742331288344,
+      "grad_norm": 37.85600478708204,
+      "learning_rate": 7.943801652892562e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3236,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 3.59375,
+      "rewards/rejected": -2.15625,
+      "step": 42
+    },
+    {
+      "epoch": 0.05276073619631902,
+      "grad_norm": 43.531972474165585,
+      "learning_rate": 7.940495867768595e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3117,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6484375,
+      "rewards/margins": 4.0625,
+      "rewards/rejected": -2.421875,
+      "step": 43
+    },
+    {
+      "epoch": 0.053987730061349694,
+      "grad_norm": 63.54372274745974,
+      "learning_rate": 7.937190082644628e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -300.0,
+      "loss": 0.4265,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 3.625,
+      "rewards/rejected": -2.53125,
+      "step": 44
+    },
+    {
+      "epoch": 0.05521472392638037,
+      "grad_norm": 42.36111410389095,
+      "learning_rate": 7.933884297520661e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3513,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 3.875,
+      "rewards/rejected": -2.640625,
+      "step": 45
+    },
+    {
+      "epoch": 0.05644171779141104,
+      "grad_norm": 44.51420354088873,
+      "learning_rate": 7.930578512396694e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3109,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -2.609375,
+      "step": 46
+    },
+    {
+      "epoch": 0.05766871165644172,
+      "grad_norm": 61.22394416060924,
+      "learning_rate": 7.927272727272727e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -368.0,
+      "loss": 0.4432,
+      "rewards/accuracies": 0.7265625,
+      "rewards/chosen": 1.2109375,
+      "rewards/margins": 3.5,
+      "rewards/rejected": -2.28125,
+      "step": 47
+    },
+    {
+      "epoch": 0.05889570552147239,
+      "grad_norm": 35.37901763029379,
+      "learning_rate": 7.92396694214876e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -352.0,
+      "logps/rejected": -282.0,
+      "loss": 0.3719,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 3.203125,
+      "rewards/rejected": -2.09375,
+      "step": 48
+    },
+    {
+      "epoch": 0.06012269938650307,
+      "grad_norm": 31.606470718574474,
+      "learning_rate": 7.920661157024793e-07,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2642,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.90625,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.75,
+      "step": 49
+    },
+    {
+      "epoch": 0.06134969325153374,
+      "grad_norm": 48.38174059914219,
+      "learning_rate": 7.917355371900826e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3848,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 3.8125,
+      "rewards/rejected": -2.359375,
+      "step": 50
+    },
+    {
+      "epoch": 0.06257668711656442,
+      "grad_norm": 41.76003291409928,
+      "learning_rate": 7.914049586776859e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -442.0,
+      "logps/rejected": -358.0,
+      "loss": 0.2936,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.890625,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.640625,
+      "step": 51
+    },
+    {
+      "epoch": 0.0638036809815951,
+      "grad_norm": 36.258140724486836,
+      "learning_rate": 7.910743801652892e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3238,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 4.21875,
+      "rewards/rejected": -2.75,
+      "step": 52
+    },
+    {
+      "epoch": 0.06503067484662577,
+      "grad_norm": 38.82725433032811,
+      "learning_rate": 7.907438016528925e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3104,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 4.03125,
+      "rewards/rejected": -2.28125,
+      "step": 53
+    },
+    {
+      "epoch": 0.06625766871165645,
+      "grad_norm": 34.79397739539489,
+      "learning_rate": 7.904132231404959e-07,
+      "logits/chosen": -0.0284423828125,
+      "logits/rejected": -0.208984375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -258.0,
+      "loss": 0.2634,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 3.78125,
+      "rewards/rejected": -2.546875,
+      "step": 54
+    },
+    {
+      "epoch": 0.06748466257668712,
+      "grad_norm": 41.093192913712485,
+      "learning_rate": 7.900826446280992e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3081,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.515625,
+      "step": 55
+    },
+    {
+      "epoch": 0.0687116564417178,
+      "grad_norm": 45.49364963721349,
+      "learning_rate": 7.897520661157025e-07,
+      "logits/chosen": -0.119140625,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3753,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 4.0625,
+      "rewards/rejected": -2.265625,
+      "step": 56
+    },
+    {
+      "epoch": 0.06993865030674846,
+      "grad_norm": 50.666379541833365,
+      "learning_rate": 7.894214876033057e-07,
+      "logits/chosen": -0.1181640625,
+      "logits/rejected": -0.203125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3742,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 3.953125,
+      "rewards/rejected": -2.421875,
+      "step": 57
+    },
+    {
+      "epoch": 0.07116564417177915,
+      "grad_norm": 47.64145593969582,
+      "learning_rate": 7.89090909090909e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.2177734375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3598,
+      "rewards/accuracies": 0.7421875,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 4.125,
+      "rewards/rejected": -2.421875,
+      "step": 58
+    },
+    {
+      "epoch": 0.07239263803680981,
+      "grad_norm": 51.11345473280857,
+      "learning_rate": 7.887603305785123e-07,
+      "logits/chosen": -0.049560546875,
+      "logits/rejected": -0.140625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -292.0,
+      "loss": 0.4133,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 3.28125,
+      "rewards/rejected": -1.9921875,
+      "step": 59
+    },
+    {
+      "epoch": 0.0736196319018405,
+      "grad_norm": 39.0647396184868,
+      "learning_rate": 7.884297520661157e-07,
+      "logits/chosen": -0.12451171875,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -284.0,
+      "loss": 0.3578,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 3.46875,
+      "rewards/rejected": -2.28125,
+      "step": 60
+    },
+    {
+      "epoch": 0.07484662576687116,
+      "grad_norm": 39.12713457741509,
+      "learning_rate": 7.88099173553719e-07,
+      "logits/chosen": -0.125,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3224,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.953125,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.578125,
+      "step": 61
+    },
+    {
+      "epoch": 0.07607361963190185,
+      "grad_norm": 41.544347269900555,
+      "learning_rate": 7.877685950413223e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -428.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3433,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 2.328125,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -2.703125,
+      "step": 62
+    },
+    {
+      "epoch": 0.07730061349693251,
+      "grad_norm": 52.28530938487766,
+      "learning_rate": 7.874380165289256e-07,
+      "logits/chosen": -0.0257568359375,
+      "logits/rejected": -0.197265625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -304.0,
+      "loss": 0.331,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 3.9375,
+      "rewards/rejected": -2.40625,
+      "step": 63
+    },
+    {
+      "epoch": 0.0785276073619632,
+      "grad_norm": 44.343811483698694,
+      "learning_rate": 7.871074380165289e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3628,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -2.5,
+      "step": 64
+    },
+    {
+      "epoch": 0.07975460122699386,
+      "grad_norm": 40.38168553746162,
+      "learning_rate": 7.867768595041322e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3193,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.828125,
+      "rewards/margins": 4.28125,
+      "rewards/rejected": -2.453125,
+      "step": 65
+    },
+    {
+      "epoch": 0.08098159509202454,
+      "grad_norm": 35.90018196646264,
+      "learning_rate": 7.864462809917355e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -318.0,
+      "logps/rejected": -260.0,
+      "loss": 0.2892,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.0,
+      "rewards/margins": 4.8125,
+      "rewards/rejected": -2.8125,
+      "step": 66
+    },
+    {
+      "epoch": 0.08220858895705521,
+      "grad_norm": 39.83708858843296,
+      "learning_rate": 7.861157024793388e-07,
+      "logits/chosen": -0.1005859375,
+      "logits/rejected": -0.224609375,
+      "logps/chosen": -332.0,
+      "logps/rejected": -294.0,
+      "loss": 0.344,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 3.90625,
+      "rewards/rejected": -2.671875,
+      "step": 67
+    },
+    {
+      "epoch": 0.0834355828220859,
+      "grad_norm": 41.25508984701749,
+      "learning_rate": 7.857851239669421e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3597,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 3.953125,
+      "rewards/rejected": -2.46875,
+      "step": 68
+    },
+    {
+      "epoch": 0.08466257668711656,
+      "grad_norm": 35.61462182457114,
+      "learning_rate": 7.854545454545454e-07,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -282.0,
+      "loss": 0.2797,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.3125,
+      "rewards/margins": 4.96875,
+      "rewards/rejected": -2.65625,
+      "step": 69
+    },
+    {
+      "epoch": 0.08588957055214724,
+      "grad_norm": 43.949799113654095,
+      "learning_rate": 7.851239669421487e-07,
+      "logits/chosen": -0.12109375,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3016,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.328125,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.34375,
+      "step": 70
+    },
+    {
+      "epoch": 0.08711656441717791,
+      "grad_norm": 40.385403541291616,
+      "learning_rate": 7.84793388429752e-07,
+      "logits/chosen": -0.068359375,
+      "logits/rejected": -0.1494140625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -278.0,
+      "loss": 0.3925,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 3.625,
+      "rewards/rejected": -2.265625,
+      "step": 71
+    },
+    {
+      "epoch": 0.08834355828220859,
+      "grad_norm": 42.66471807743761,
+      "learning_rate": 7.844628099173554e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -430.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2955,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9609375,
+      "rewards/margins": 4.25,
+      "rewards/rejected": -2.28125,
+      "step": 72
+    },
+    {
+      "epoch": 0.08957055214723926,
+      "grad_norm": 45.57465316917817,
+      "learning_rate": 7.841322314049587e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -348.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3924,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.5625,
+      "step": 73
+    },
+    {
+      "epoch": 0.09079754601226994,
+      "grad_norm": 41.41678309244041,
+      "learning_rate": 7.838016528925619e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2911,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 3.9375,
+      "rewards/rejected": -2.65625,
+      "step": 74
+    },
+    {
+      "epoch": 0.09202453987730061,
+      "grad_norm": 36.36740264063768,
+      "learning_rate": 7.834710743801652e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -318.0,
+      "loss": 0.271,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.703125,
+      "step": 75
+    },
+    {
+      "epoch": 0.09325153374233129,
+      "grad_norm": 51.152036733559505,
+      "learning_rate": 7.831404958677685e-07,
+      "logits/chosen": -0.1513671875,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -288.0,
+      "loss": 0.3262,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -2.53125,
+      "step": 76
+    },
+    {
+      "epoch": 0.09447852760736196,
+      "grad_norm": 34.55182641721203,
+      "learning_rate": 7.828099173553718e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2774,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.4453125,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -2.875,
+      "step": 77
+    },
+    {
+      "epoch": 0.09570552147239264,
+      "grad_norm": 40.758007342891354,
+      "learning_rate": 7.824793388429752e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2953,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -2.859375,
+      "step": 78
+    },
+    {
+      "epoch": 0.09693251533742331,
+      "grad_norm": 40.243771187807226,
+      "learning_rate": 7.821487603305785e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -286.0,
+      "loss": 0.324,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.796875,
+      "step": 79
+    },
+    {
+      "epoch": 0.09815950920245399,
+      "grad_norm": 37.86169713828253,
+      "learning_rate": 7.818181818181818e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2795,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.0,
+      "step": 80
+    },
+    {
+      "epoch": 0.09938650306748466,
+      "grad_norm": 36.69542578194377,
+      "learning_rate": 7.814876033057851e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -436.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.9375,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.09375,
+      "step": 81
+    },
+    {
+      "epoch": 0.10061349693251534,
+      "grad_norm": 52.318975452049024,
+      "learning_rate": 7.811570247933884e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3562,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -3.1875,
+      "step": 82
+    },
+    {
+      "epoch": 0.10184049079754601,
+      "grad_norm": 48.36250064527517,
+      "learning_rate": 7.808264462809916e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3065,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 4.5625,
+      "rewards/rejected": -3.234375,
+      "step": 83
+    },
+    {
+      "epoch": 0.10306748466257669,
+      "grad_norm": 56.4532578365936,
+      "learning_rate": 7.80495867768595e-07,
+      "logits/chosen": -0.11083984375,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -448.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2942,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -3.53125,
+      "step": 84
+    },
+    {
+      "epoch": 0.10429447852760736,
+      "grad_norm": 40.84421229525563,
+      "learning_rate": 7.801652892561983e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3021,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.484375,
+      "step": 85
+    },
+    {
+      "epoch": 0.10552147239263804,
+      "grad_norm": 42.8824866796569,
+      "learning_rate": 7.798347107438016e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -302.0,
+      "loss": 0.3346,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.03125,
+      "step": 86
+    },
+    {
+      "epoch": 0.1067484662576687,
+      "grad_norm": 32.177789360920215,
+      "learning_rate": 7.795041322314049e-07,
+      "logits/chosen": -0.17578125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -260.0,
+      "loss": 0.2835,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 4.25,
+      "rewards/rejected": -3.46875,
+      "step": 87
+    },
+    {
+      "epoch": 0.10797546012269939,
+      "grad_norm": 34.28909140331474,
+      "learning_rate": 7.791735537190082e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -276.0,
+      "loss": 0.265,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -3.3125,
+      "step": 88
+    },
+    {
+      "epoch": 0.10920245398773006,
+      "grad_norm": 42.40518360379319,
+      "learning_rate": 7.788429752066116e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3464,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.4140625,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -3.09375,
+      "step": 89
+    },
+    {
+      "epoch": 0.11042944785276074,
+      "grad_norm": 49.53932371851528,
+      "learning_rate": 7.785123966942149e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3253,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.98828125,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -3.3125,
+      "step": 90
+    },
+    {
+      "epoch": 0.1116564417177914,
+      "grad_norm": 40.48862678893552,
+      "learning_rate": 7.781818181818182e-07,
+      "logits/chosen": -0.039794921875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2845,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -3.203125,
+      "step": 91
+    },
+    {
+      "epoch": 0.11288343558282209,
+      "grad_norm": 44.87681376045074,
+      "learning_rate": 7.778512396694214e-07,
+      "logits/chosen": -0.126953125,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2738,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.5,
+      "step": 92
+    },
+    {
+      "epoch": 0.11411042944785275,
+      "grad_norm": 36.61224632474363,
+      "learning_rate": 7.775206611570247e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -300.0,
+      "loss": 0.282,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.3125,
+      "step": 93
+    },
+    {
+      "epoch": 0.11533742331288344,
+      "grad_norm": 42.24461574764279,
+      "learning_rate": 7.77190082644628e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2813,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.234375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.140625,
+      "step": 94
+    },
+    {
+      "epoch": 0.1165644171779141,
+      "grad_norm": 42.13498046583409,
+      "learning_rate": 7.768595041322314e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.7734375,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -3.390625,
+      "step": 95
+    },
+    {
+      "epoch": 0.11779141104294479,
+      "grad_norm": 39.13305361039425,
+      "learning_rate": 7.765289256198347e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -310.0,
+      "loss": 0.253,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.203125,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -3.40625,
+      "step": 96
+    },
+    {
+      "epoch": 0.11901840490797547,
+      "grad_norm": 46.1275059887527,
+      "learning_rate": 7.76198347107438e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -352.0,
+      "loss": 0.357,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -2.890625,
+      "step": 97
+    },
+    {
+      "epoch": 0.12024539877300613,
+      "grad_norm": 38.580559184520915,
+      "learning_rate": 7.758677685950413e-07,
+      "logits/chosen": -0.041015625,
+      "logits/rejected": -0.248046875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -282.0,
+      "loss": 0.3015,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.828125,
+      "step": 98
+    },
+    {
+      "epoch": 0.12147239263803682,
+      "grad_norm": 38.975117366609354,
+      "learning_rate": 7.755371900826446e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -340.0,
+      "logps/rejected": -244.0,
+      "loss": 0.2586,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.84375,
+      "step": 99
+    },
+    {
+      "epoch": 0.12269938650306748,
+      "grad_norm": 38.2314125667712,
+      "learning_rate": 7.752066115702479e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3048,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 4.875,
+      "rewards/rejected": -3.09375,
+      "step": 100
+    },
+    {
+      "epoch": 0.12392638036809817,
+      "grad_norm": 47.72088602593496,
+      "learning_rate": 7.748760330578512e-07,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3549,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.921875,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -2.984375,
+      "step": 101
+    },
+    {
+      "epoch": 0.12515337423312883,
+      "grad_norm": 39.20874470259191,
+      "learning_rate": 7.745454545454545e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3073,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.9375,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.71875,
+      "step": 102
+    },
+    {
+      "epoch": 0.1263803680981595,
+      "grad_norm": 36.963577318972284,
+      "learning_rate": 7.742148760330578e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2878,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7265625,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -2.75,
+      "step": 103
+    },
+    {
+      "epoch": 0.1276073619631902,
+      "grad_norm": 50.89003123288155,
+      "learning_rate": 7.738842975206611e-07,
+      "logits/chosen": -0.107421875,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3679,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 4.40625,
+      "rewards/rejected": -2.71875,
+      "step": 104
+    },
+    {
+      "epoch": 0.12883435582822086,
+      "grad_norm": 34.19259318441508,
+      "learning_rate": 7.735537190082644e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2586,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.984375,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.140625,
+      "step": 105
+    },
+    {
+      "epoch": 0.13006134969325153,
+      "grad_norm": 38.48374397973381,
+      "learning_rate": 7.732231404958677e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3111,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.9140625,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -2.828125,
+      "step": 106
+    },
+    {
+      "epoch": 0.1312883435582822,
+      "grad_norm": 41.75637341421824,
+      "learning_rate": 7.728925619834711e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2869,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.84375,
+      "step": 107
+    },
+    {
+      "epoch": 0.1325153374233129,
+      "grad_norm": 33.954750583294235,
+      "learning_rate": 7.725619834710744e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2625,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.328125,
+      "step": 108
+    },
+    {
+      "epoch": 0.13374233128834356,
+      "grad_norm": 36.44619320062165,
+      "learning_rate": 7.722314049586777e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2958,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.6875,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -2.96875,
+      "step": 109
+    },
+    {
+      "epoch": 0.13496932515337423,
+      "grad_norm": 41.465536696868014,
+      "learning_rate": 7.719008264462809e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3213,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 2.234375,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.015625,
+      "step": 110
+    },
+    {
+      "epoch": 0.1361963190184049,
+      "grad_norm": 34.97820982991418,
+      "learning_rate": 7.715702479338842e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -330.0,
+      "logps/rejected": -268.0,
+      "loss": 0.2777,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 4.40625,
+      "rewards/rejected": -2.75,
+      "step": 111
+    },
+    {
+      "epoch": 0.1374233128834356,
+      "grad_norm": 37.33233370442031,
+      "learning_rate": 7.712396694214875e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -340.0,
+      "logps/rejected": -278.0,
+      "loss": 0.2729,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 4.6875,
+      "rewards/rejected": -3.125,
+      "step": 112
+    },
+    {
+      "epoch": 0.13865030674846626,
+      "grad_norm": 39.32831836749558,
+      "learning_rate": 7.709090909090909e-07,
+      "logits/chosen": -0.1044921875,
+      "logits/rejected": -0.2001953125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -302.0,
+      "loss": 0.3238,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.84375,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -2.890625,
+      "step": 113
+    },
+    {
+      "epoch": 0.13987730061349693,
+      "grad_norm": 43.77160490529073,
+      "learning_rate": 7.705785123966942e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3356,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -2.96875,
+      "step": 114
+    },
+    {
+      "epoch": 0.1411042944785276,
+      "grad_norm": 45.86173030028754,
+      "learning_rate": 7.702479338842975e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3666,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 4.0,
+      "rewards/rejected": -2.96875,
+      "step": 115
+    },
+    {
+      "epoch": 0.1423312883435583,
+      "grad_norm": 31.833834890400677,
+      "learning_rate": 7.699173553719008e-07,
+      "logits/chosen": -0.2001953125,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2561,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.078125,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.140625,
+      "step": 116
+    },
+    {
+      "epoch": 0.14355828220858896,
+      "grad_norm": 34.65342816309822,
+      "learning_rate": 7.695867768595041e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -324.0,
+      "logps/rejected": -262.0,
+      "loss": 0.2802,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -3.453125,
+      "step": 117
+    },
+    {
+      "epoch": 0.14478527607361963,
+      "grad_norm": 40.64680750199124,
+      "learning_rate": 7.692561983471075e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3232,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.09375,
+      "step": 118
+    },
+    {
+      "epoch": 0.1460122699386503,
+      "grad_norm": 39.06542225750342,
+      "learning_rate": 7.689256198347107e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3009,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -3.34375,
+      "step": 119
+    },
+    {
+      "epoch": 0.147239263803681,
+      "grad_norm": 45.793161618342445,
+      "learning_rate": 7.68595041322314e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2967,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 2.421875,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -3.234375,
+      "step": 120
+    },
+    {
+      "epoch": 0.14846625766871166,
+      "grad_norm": 36.35506762175528,
+      "learning_rate": 7.682644628099173e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2827,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 2.265625,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -3.09375,
+      "step": 121
+    },
+    {
+      "epoch": 0.14969325153374233,
+      "grad_norm": 43.61172830028863,
+      "learning_rate": 7.679338842975206e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2954,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -3.015625,
+      "step": 122
+    },
+    {
+      "epoch": 0.150920245398773,
+      "grad_norm": 44.76510892093205,
+      "learning_rate": 7.676033057851239e-07,
+      "logits/chosen": -0.054443359375,
+      "logits/rejected": -0.2294921875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -278.0,
+      "loss": 0.3266,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 4.40625,
+      "rewards/rejected": -3.171875,
+      "step": 123
+    },
+    {
+      "epoch": 0.1521472392638037,
+      "grad_norm": 34.74558330134135,
+      "learning_rate": 7.672727272727273e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -350.0,
+      "logps/rejected": -280.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -3.46875,
+      "step": 124
+    },
+    {
+      "epoch": 0.15337423312883436,
+      "grad_norm": 42.40382525839625,
+      "learning_rate": 7.669421487603306e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2816,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.765625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.5625,
+      "step": 125
+    },
+    {
+      "epoch": 0.15460122699386503,
+      "grad_norm": 54.09847924655591,
+      "learning_rate": 7.666115702479339e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -302.0,
+      "loss": 0.3503,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 4.34375,
+      "rewards/rejected": -3.34375,
+      "step": 126
+    },
+    {
+      "epoch": 0.1558282208588957,
+      "grad_norm": 42.126209906017394,
+      "learning_rate": 7.662809917355372e-07,
+      "logits/chosen": -0.09375,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3524,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.96875,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.15625,
+      "step": 127
+    },
+    {
+      "epoch": 0.1570552147239264,
+      "grad_norm": 41.454801423663085,
+      "learning_rate": 7.659504132231404e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2841,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8671875,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.390625,
+      "step": 128
+    },
+    {
+      "epoch": 0.15828220858895706,
+      "grad_norm": 37.48031578456685,
+      "learning_rate": 7.656198347107437e-07,
+      "logits/chosen": -0.0478515625,
+      "logits/rejected": -0.1796875,
+      "logps/chosen": -336.0,
+      "logps/rejected": -284.0,
+      "loss": 0.3245,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -3.0,
+      "step": 129
+    },
+    {
+      "epoch": 0.15950920245398773,
+      "grad_norm": 38.14450880210714,
+      "learning_rate": 7.65289256198347e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3273,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -3.421875,
+      "step": 130
+    },
+    {
+      "epoch": 0.1607361963190184,
+      "grad_norm": 35.19502262660919,
+      "learning_rate": 7.649586776859504e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2106,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.953125,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.59375,
+      "step": 131
+    },
+    {
+      "epoch": 0.1619631901840491,
+      "grad_norm": 42.07184086801236,
+      "learning_rate": 7.646280991735537e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3252,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.765625,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -3.40625,
+      "step": 132
+    },
+    {
+      "epoch": 0.16319018404907976,
+      "grad_norm": 43.059635351413924,
+      "learning_rate": 7.64297520661157e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2872,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.703125,
+      "step": 133
+    },
+    {
+      "epoch": 0.16441717791411042,
+      "grad_norm": 48.09563853008669,
+      "learning_rate": 7.639669421487603e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -450.0,
+      "logps/rejected": -358.0,
+      "loss": 0.2946,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 4.625,
+      "rewards/rejected": -3.109375,
+      "step": 134
+    },
+    {
+      "epoch": 0.1656441717791411,
+      "grad_norm": 46.36508764131004,
+      "learning_rate": 7.636363636363636e-07,
+      "logits/chosen": -0.05712890625,
+      "logits/rejected": -0.171875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3654,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.9765625,
+      "rewards/margins": 4.375,
+      "rewards/rejected": -3.390625,
+      "step": 135
+    },
+    {
+      "epoch": 0.1668711656441718,
+      "grad_norm": 39.349987951574626,
+      "learning_rate": 7.63305785123967e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3195,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -3.78125,
+      "step": 136
+    },
+    {
+      "epoch": 0.16809815950920245,
+      "grad_norm": 46.78307320347251,
+      "learning_rate": 7.629752066115702e-07,
+      "logits/chosen": -0.061279296875,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3621,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 4.25,
+      "rewards/rejected": -3.078125,
+      "step": 137
+    },
+    {
+      "epoch": 0.16932515337423312,
+      "grad_norm": 37.18351460769713,
+      "learning_rate": 7.626446280991735e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3022,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.46875,
+      "step": 138
+    },
+    {
+      "epoch": 0.1705521472392638,
+      "grad_norm": 42.545002180919454,
+      "learning_rate": 7.623140495867768e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -362.0,
+      "loss": 0.352,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -3.828125,
+      "step": 139
+    },
+    {
+      "epoch": 0.17177914110429449,
+      "grad_norm": 47.45496980791103,
+      "learning_rate": 7.619834710743801e-07,
+      "logits/chosen": -0.10400390625,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3166,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.5625,
+      "step": 140
+    },
+    {
+      "epoch": 0.17300613496932515,
+      "grad_norm": 34.622965144928166,
+      "learning_rate": 7.616528925619834e-07,
+      "logits/chosen": -0.0849609375,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.251,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.515625,
+      "step": 141
+    },
+    {
+      "epoch": 0.17423312883435582,
+      "grad_norm": 40.37071693879945,
+      "learning_rate": 7.613223140495868e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3233,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.640625,
+      "step": 142
+    },
+    {
+      "epoch": 0.1754601226993865,
+      "grad_norm": 41.48963188375289,
+      "learning_rate": 7.609917355371901e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2941,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.828125,
+      "step": 143
+    },
+    {
+      "epoch": 0.17668711656441718,
+      "grad_norm": 40.697834722668915,
+      "learning_rate": 7.606611570247934e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2905,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.96875,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.59375,
+      "step": 144
+    },
+    {
+      "epoch": 0.17791411042944785,
+      "grad_norm": 39.56968971060715,
+      "learning_rate": 7.603305785123967e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2715,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.65625,
+      "step": 145
+    },
+    {
+      "epoch": 0.17914110429447852,
+      "grad_norm": 34.52865223755899,
+      "learning_rate": 7.599999999999999e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2436,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.78125,
+      "step": 146
+    },
+    {
+      "epoch": 0.18036809815950922,
+      "grad_norm": 35.86987627703754,
+      "learning_rate": 7.596694214876032e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2741,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.84375,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.71875,
+      "step": 147
+    },
+    {
+      "epoch": 0.18159509202453988,
+      "grad_norm": 49.681572441729514,
+      "learning_rate": 7.593388429752066e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3761,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.625,
+      "step": 148
+    },
+    {
+      "epoch": 0.18282208588957055,
+      "grad_norm": 43.932202256712266,
+      "learning_rate": 7.590082644628099e-07,
+      "logits/chosen": -0.11474609375,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2534,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.984375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.390625,
+      "step": 149
+    },
+    {
+      "epoch": 0.18404907975460122,
+      "grad_norm": 39.402366627439726,
+      "learning_rate": 7.586776859504132e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3158,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -3.09375,
+      "step": 150
+    },
+    {
+      "epoch": 0.18527607361963191,
+      "grad_norm": 41.79252221918161,
+      "learning_rate": 7.583471074380165e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3211,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -3.03125,
+      "step": 151
+    },
+    {
+      "epoch": 0.18650306748466258,
+      "grad_norm": 35.715508495832665,
+      "learning_rate": 7.580165289256198e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -298.0,
+      "loss": 0.2622,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.921875,
+      "step": 152
+    },
+    {
+      "epoch": 0.18773006134969325,
+      "grad_norm": 36.417969122212845,
+      "learning_rate": 7.576859504132231e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2871,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.625,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.390625,
+      "step": 153
+    },
+    {
+      "epoch": 0.18895705521472392,
+      "grad_norm": 45.428337431051204,
+      "learning_rate": 7.573553719008265e-07,
+      "logits/chosen": -0.09716796875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3165,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 2.046875,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -2.984375,
+      "step": 154
+    },
+    {
+      "epoch": 0.1901840490797546,
+      "grad_norm": 28.34310543484037,
+      "learning_rate": 7.570247933884297e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1864,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 2.5625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -3.65625,
+      "step": 155
+    },
+    {
+      "epoch": 0.19141104294478528,
+      "grad_norm": 43.05423041815277,
+      "learning_rate": 7.56694214876033e-07,
+      "logits/chosen": -0.11865234375,
+      "logits/rejected": -0.2314453125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2995,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 4.5625,
+      "rewards/rejected": -3.09375,
+      "step": 156
+    },
+    {
+      "epoch": 0.19263803680981595,
+      "grad_norm": 35.31331641855709,
+      "learning_rate": 7.563636363636363e-07,
+      "logits/chosen": -0.11474609375,
+      "logits/rejected": -0.244140625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2791,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 4.875,
+      "rewards/rejected": -3.328125,
+      "step": 157
+    },
+    {
+      "epoch": 0.19386503067484662,
+      "grad_norm": 32.52885612180771,
+      "learning_rate": 7.560330578512396e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2615,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.6875,
+      "step": 158
+    },
+    {
+      "epoch": 0.1950920245398773,
+      "grad_norm": 40.74250145862389,
+      "learning_rate": 7.55702479338843e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -318.0,
+      "loss": 0.312,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.25,
+      "step": 159
+    },
+    {
+      "epoch": 0.19631901840490798,
+      "grad_norm": 32.03909927740134,
+      "learning_rate": 7.553719008264463e-07,
+      "logits/chosen": -0.10009765625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2149,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.765625,
+      "step": 160
+    },
+    {
+      "epoch": 0.19754601226993865,
+      "grad_norm": 33.48229976640365,
+      "learning_rate": 7.550413223140496e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2313,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.8984375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -3.703125,
+      "step": 161
+    },
+    {
+      "epoch": 0.19877300613496932,
+      "grad_norm": 44.9989690142436,
+      "learning_rate": 7.547107438016529e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3547,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.390625,
+      "step": 162
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 45.392211600179095,
+      "learning_rate": 7.543801652892561e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3203,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.4375,
+      "step": 163
+    },
+    {
+      "epoch": 0.20122699386503068,
+      "grad_norm": 30.10144812491185,
+      "learning_rate": 7.540495867768594e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -298.0,
+      "loss": 0.2205,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -3.953125,
+      "step": 164
+    },
+    {
+      "epoch": 0.20245398773006135,
+      "grad_norm": 41.70055038557836,
+      "learning_rate": 7.537190082644627e-07,
+      "logits/chosen": -0.078125,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -440.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2684,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.7578125,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.484375,
+      "step": 165
+    },
+    {
+      "epoch": 0.20368098159509201,
+      "grad_norm": 40.19134249676154,
+      "learning_rate": 7.533884297520661e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3382,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -3.296875,
+      "step": 166
+    },
+    {
+      "epoch": 0.2049079754601227,
+      "grad_norm": 40.73221481198693,
+      "learning_rate": 7.530578512396694e-07,
+      "logits/chosen": -0.1240234375,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3475,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 2.21875,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -3.40625,
+      "step": 167
+    },
+    {
+      "epoch": 0.20613496932515338,
+      "grad_norm": 36.56704570459498,
+      "learning_rate": 7.527272727272727e-07,
+      "logits/chosen": -0.10888671875,
+      "logits/rejected": -0.2236328125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -276.0,
+      "loss": 0.2594,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.0,
+      "step": 168
+    },
+    {
+      "epoch": 0.20736196319018405,
+      "grad_norm": 36.86789794879476,
+      "learning_rate": 7.52396694214876e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2793,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.75,
+      "step": 169
+    },
+    {
+      "epoch": 0.2085889570552147,
+      "grad_norm": 43.700977925857806,
+      "learning_rate": 7.520661157024793e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -448.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2994,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.046875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -3.765625,
+      "step": 170
+    },
+    {
+      "epoch": 0.2098159509202454,
+      "grad_norm": 33.49786387415663,
+      "learning_rate": 7.517355371900827e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2804,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.7734375,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.0625,
+      "step": 171
+    },
+    {
+      "epoch": 0.21104294478527608,
+      "grad_norm": 36.775839118353844,
+      "learning_rate": 7.514049586776859e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -436.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2514,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.03125,
+      "step": 172
+    },
+    {
+      "epoch": 0.21226993865030674,
+      "grad_norm": 29.010747243935523,
+      "learning_rate": 7.510743801652892e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2075,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.296875,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.125,
+      "step": 173
+    },
+    {
+      "epoch": 0.2134969325153374,
+      "grad_norm": 42.10282379400345,
+      "learning_rate": 7.507438016528925e-07,
+      "logits/chosen": -0.035400390625,
+      "logits/rejected": -0.248046875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3371,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.890625,
+      "step": 174
+    },
+    {
+      "epoch": 0.2147239263803681,
+      "grad_norm": 44.4053310320228,
+      "learning_rate": 7.504132231404958e-07,
+      "logits/chosen": -0.17578125,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3132,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.9609375,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -3.921875,
+      "step": 175
+    },
+    {
+      "epoch": 0.21595092024539878,
+      "grad_norm": 38.79125305740806,
+      "learning_rate": 7.500826446280991e-07,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -354.0,
+      "loss": 0.315,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -3.875,
+      "step": 176
+    },
+    {
+      "epoch": 0.21717791411042944,
+      "grad_norm": 29.08349894872895,
+      "learning_rate": 7.497520661157025e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2249,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.453125,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -4.375,
+      "step": 177
+    },
+    {
+      "epoch": 0.2184049079754601,
+      "grad_norm": 39.84454997153339,
+      "learning_rate": 7.494214876033058e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3147,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.375,
+      "step": 178
+    },
+    {
+      "epoch": 0.2196319018404908,
+      "grad_norm": 32.49552053062871,
+      "learning_rate": 7.490909090909091e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2585,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.34375,
+      "step": 179
+    },
+    {
+      "epoch": 0.22085889570552147,
+      "grad_norm": 41.53796098392134,
+      "learning_rate": 7.487603305785124e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3355,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.75,
+      "step": 180
+    },
+    {
+      "epoch": 0.22208588957055214,
+      "grad_norm": 43.27113784190138,
+      "learning_rate": 7.484297520661156e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -372.0,
+      "loss": 0.3558,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.03125,
+      "step": 181
+    },
+    {
+      "epoch": 0.2233128834355828,
+      "grad_norm": 40.060092182372635,
+      "learning_rate": 7.480991735537189e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3329,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.92578125,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -4.0625,
+      "step": 182
+    },
+    {
+      "epoch": 0.2245398773006135,
+      "grad_norm": 42.893284410160845,
+      "learning_rate": 7.477685950413223e-07,
+      "logits/chosen": -0.0888671875,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -374.0,
+      "loss": 0.3303,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -3.703125,
+      "step": 183
+    },
+    {
+      "epoch": 0.22576687116564417,
+      "grad_norm": 44.81738203832481,
+      "learning_rate": 7.474380165289256e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.2431640625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3382,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.0625,
+      "step": 184
+    },
+    {
+      "epoch": 0.22699386503067484,
+      "grad_norm": 49.18179746490089,
+      "learning_rate": 7.471074380165289e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3425,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -3.734375,
+      "step": 185
+    },
+    {
+      "epoch": 0.2282208588957055,
+      "grad_norm": 38.62632364463619,
+      "learning_rate": 7.467768595041322e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3061,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.953125,
+      "step": 186
+    },
+    {
+      "epoch": 0.2294478527607362,
+      "grad_norm": 37.48238102625356,
+      "learning_rate": 7.464462809917355e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2719,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.375,
+      "step": 187
+    },
+    {
+      "epoch": 0.23067484662576687,
+      "grad_norm": 33.9965710800535,
+      "learning_rate": 7.461157024793388e-07,
+      "logits/chosen": -0.0712890625,
+      "logits/rejected": -0.1689453125,
+      "logps/chosen": -332.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2908,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.09375,
+      "step": 188
+    },
+    {
+      "epoch": 0.23190184049079754,
+      "grad_norm": 32.6251624460159,
+      "learning_rate": 7.457851239669422e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2196,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.3125,
+      "step": 189
+    },
+    {
+      "epoch": 0.2331288343558282,
+      "grad_norm": 33.86427029616883,
+      "learning_rate": 7.454545454545454e-07,
+      "logits/chosen": -0.05810546875,
+      "logits/rejected": -0.224609375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2859,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.875,
+      "step": 190
+    },
+    {
+      "epoch": 0.2343558282208589,
+      "grad_norm": 48.95754374693823,
+      "learning_rate": 7.451239669421487e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3495,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.1875,
+      "step": 191
+    },
+    {
+      "epoch": 0.23558282208588957,
+      "grad_norm": 44.694192367882906,
+      "learning_rate": 7.44793388429752e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -292.0,
+      "loss": 0.3014,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.578125,
+      "step": 192
+    },
+    {
+      "epoch": 0.23680981595092024,
+      "grad_norm": 34.25684662561766,
+      "learning_rate": 7.444628099173553e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2177,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.28125,
+      "step": 193
+    },
+    {
+      "epoch": 0.23803680981595093,
+      "grad_norm": 39.119509435238996,
+      "learning_rate": 7.441322314049586e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2953,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.4375,
+      "step": 194
+    },
+    {
+      "epoch": 0.2392638036809816,
+      "grad_norm": 34.817335787332766,
+      "learning_rate": 7.43801652892562e-07,
+      "logits/chosen": -0.1181640625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2301,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.4375,
+      "step": 195
+    },
+    {
+      "epoch": 0.24049079754601227,
+      "grad_norm": 40.102802913255154,
+      "learning_rate": 7.434710743801653e-07,
+      "logits/chosen": -0.09130859375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -290.0,
+      "loss": 0.297,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -3.828125,
+      "step": 196
+    },
+    {
+      "epoch": 0.24171779141104294,
+      "grad_norm": 34.798967486094014,
+      "learning_rate": 7.431404958677686e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2803,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.0625,
+      "step": 197
+    },
+    {
+      "epoch": 0.24294478527607363,
+      "grad_norm": 39.54376141056963,
+      "learning_rate": 7.428099173553719e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.2265625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -258.0,
+      "loss": 0.3347,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.671875,
+      "step": 198
+    },
+    {
+      "epoch": 0.2441717791411043,
+      "grad_norm": 29.961664099131877,
+      "learning_rate": 7.424793388429751e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -428.0,
+      "logps/rejected": -310.0,
+      "loss": 0.1903,
+      "rewards/accuracies": 0.9140625,
+      "rewards/chosen": 2.359375,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.09375,
+      "step": 199
+    },
+    {
+      "epoch": 0.24539877300613497,
+      "grad_norm": 41.86791512110987,
+      "learning_rate": 7.421487603305784e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -370.0,
+      "loss": 0.3056,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.203125,
+      "rewards/margins": 6.125,
+      "rewards/rejected": -3.921875,
+      "step": 200
+    },
+    {
+      "epoch": 0.24662576687116564,
+      "grad_norm": 32.0817049971399,
+      "learning_rate": 7.418181818181818e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2166,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 2.0625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -3.734375,
+      "step": 201
+    },
+    {
+      "epoch": 0.24785276073619633,
+      "grad_norm": 35.7245529395789,
+      "learning_rate": 7.414876033057851e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2641,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -3.640625,
+      "step": 202
+    },
+    {
+      "epoch": 0.249079754601227,
+      "grad_norm": 42.27175209208064,
+      "learning_rate": 7.411570247933884e-07,
+      "logits/chosen": -0.08935546875,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4055,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8828125,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.03125,
+      "step": 203
+    },
+    {
+      "epoch": 0.25030674846625767,
+      "grad_norm": 39.50994905620242,
+      "learning_rate": 7.408264462809917e-07,
+      "logits/chosen": -0.0771484375,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -268.0,
+      "loss": 0.3395,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -3.671875,
+      "step": 204
+    },
+    {
+      "epoch": 0.25153374233128833,
+      "grad_norm": 37.49332179023069,
+      "learning_rate": 7.40495867768595e-07,
+      "logits/chosen": -0.02880859375,
+      "logits/rejected": -0.173828125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3209,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.84375,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.265625,
+      "step": 205
+    },
+    {
+      "epoch": 0.252760736196319,
+      "grad_norm": 27.399492756017406,
+      "learning_rate": 7.401652892561984e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2051,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.28125,
+      "step": 206
+    },
+    {
+      "epoch": 0.25398773006134967,
+      "grad_norm": 43.41671912422527,
+      "learning_rate": 7.398347107438017e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3584,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 2.140625,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -3.71875,
+      "step": 207
+    },
+    {
+      "epoch": 0.2552147239263804,
+      "grad_norm": 38.06892744188352,
+      "learning_rate": 7.395041322314049e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3179,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.71875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.09375,
+      "step": 208
+    },
+    {
+      "epoch": 0.25644171779141106,
+      "grad_norm": 38.21598347000612,
+      "learning_rate": 7.391735537190082e-07,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -336.0,
+      "loss": 0.306,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -3.90625,
+      "step": 209
+    },
+    {
+      "epoch": 0.25766871165644173,
+      "grad_norm": 45.46016539899385,
+      "learning_rate": 7.388429752066115e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -364.0,
+      "loss": 0.3476,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.9375,
+      "step": 210
+    },
+    {
+      "epoch": 0.2588957055214724,
+      "grad_norm": 38.3264065355804,
+      "learning_rate": 7.385123966942148e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3035,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.53125,
+      "step": 211
+    },
+    {
+      "epoch": 0.26012269938650306,
+      "grad_norm": 37.36895578645757,
+      "learning_rate": 7.381818181818182e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2805,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.28125,
+      "step": 212
+    },
+    {
+      "epoch": 0.26134969325153373,
+      "grad_norm": 33.79330367194955,
+      "learning_rate": 7.378512396694215e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -284.0,
+      "loss": 0.2588,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.9375,
+      "step": 213
+    },
+    {
+      "epoch": 0.2625766871165644,
+      "grad_norm": 41.91194027673872,
+      "learning_rate": 7.375206611570248e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3085,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -4.125,
+      "step": 214
+    },
+    {
+      "epoch": 0.26380368098159507,
+      "grad_norm": 42.67216156371333,
+      "learning_rate": 7.371900826446281e-07,
+      "logits/chosen": -0.0927734375,
+      "logits/rejected": -0.1953125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3066,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 4.96875,
+      "rewards/rejected": -3.8125,
+      "step": 215
+    },
+    {
+      "epoch": 0.2650306748466258,
+      "grad_norm": 41.6917127385005,
+      "learning_rate": 7.368595041322314e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -330.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3077,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -3.734375,
+      "step": 216
+    },
+    {
+      "epoch": 0.26625766871165646,
+      "grad_norm": 41.999916191643734,
+      "learning_rate": 7.365289256198346e-07,
+      "logits/chosen": -0.05322265625,
+      "logits/rejected": -0.20703125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3621,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.421875,
+      "step": 217
+    },
+    {
+      "epoch": 0.2674846625766871,
+      "grad_norm": 48.35045291177634,
+      "learning_rate": 7.36198347107438e-07,
+      "logits/chosen": -0.10009765625,
+      "logits/rejected": -0.21875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3724,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 4.875,
+      "rewards/rejected": -3.515625,
+      "step": 218
+    },
+    {
+      "epoch": 0.2687116564417178,
+      "grad_norm": 36.99776737120032,
+      "learning_rate": 7.358677685950413e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3041,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.4140625,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -3.875,
+      "step": 219
+    },
+    {
+      "epoch": 0.26993865030674846,
+      "grad_norm": 49.4734461720144,
+      "learning_rate": 7.355371900826446e-07,
+      "logits/chosen": -0.05078125,
+      "logits/rejected": -0.1318359375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3781,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 4.375,
+      "rewards/rejected": -3.421875,
+      "step": 220
+    },
+    {
+      "epoch": 0.27116564417177913,
+      "grad_norm": 50.14968062203286,
+      "learning_rate": 7.352066115702479e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -292.0,
+      "loss": 0.3828,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.796875,
+      "step": 221
+    },
+    {
+      "epoch": 0.2723926380368098,
+      "grad_norm": 35.95574239692987,
+      "learning_rate": 7.348760330578512e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2649,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.09375,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.15625,
+      "step": 222
+    },
+    {
+      "epoch": 0.27361963190184047,
+      "grad_norm": 40.443295860960646,
+      "learning_rate": 7.345454545454545e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2888,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.75,
+      "step": 223
+    },
+    {
+      "epoch": 0.2748466257668712,
+      "grad_norm": 42.19474227390439,
+      "learning_rate": 7.342148760330579e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -446.0,
+      "logps/rejected": -334.0,
+      "loss": 0.288,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.09375,
+      "step": 224
+    },
+    {
+      "epoch": 0.27607361963190186,
+      "grad_norm": 37.9965018970918,
+      "learning_rate": 7.338842975206612e-07,
+      "logits/chosen": -0.10546875,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -336.0,
+      "logps/rejected": -304.0,
+      "loss": 0.261,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.40625,
+      "step": 225
+    },
+    {
+      "epoch": 0.2773006134969325,
+      "grad_norm": 38.80764018954137,
+      "learning_rate": 7.335537190082644e-07,
+      "logits/chosen": -0.119140625,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3311,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -3.890625,
+      "step": 226
+    },
+    {
+      "epoch": 0.2785276073619632,
+      "grad_norm": 42.55177823975192,
+      "learning_rate": 7.332231404958677e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3198,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.40625,
+      "step": 227
+    },
+    {
+      "epoch": 0.27975460122699386,
+      "grad_norm": 47.83061041994659,
+      "learning_rate": 7.32892561983471e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3258,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -4.125,
+      "step": 228
+    },
+    {
+      "epoch": 0.2809815950920245,
+      "grad_norm": 36.471929966124165,
+      "learning_rate": 7.325619834710743e-07,
+      "logits/chosen": -0.07373046875,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2558,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.59375,
+      "step": 229
+    },
+    {
+      "epoch": 0.2822085889570552,
+      "grad_norm": 36.373451110997884,
+      "learning_rate": 7.322314049586777e-07,
+      "logits/chosen": -0.06884765625,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2454,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.28125,
+      "step": 230
+    },
+    {
+      "epoch": 0.28343558282208586,
+      "grad_norm": 44.837599338102365,
+      "learning_rate": 7.31900826446281e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3895,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.625,
+      "step": 231
+    },
+    {
+      "epoch": 0.2846625766871166,
+      "grad_norm": 40.26904672106654,
+      "learning_rate": 7.315702479338843e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3097,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -4.0625,
+      "step": 232
+    },
+    {
+      "epoch": 0.28588957055214725,
+      "grad_norm": 37.854698103189165,
+      "learning_rate": 7.312396694214876e-07,
+      "logits/chosen": -0.10888671875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -288.0,
+      "loss": 0.3055,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.0625,
+      "step": 233
+    },
+    {
+      "epoch": 0.2871165644171779,
+      "grad_norm": 38.018318477115464,
+      "learning_rate": 7.309090909090909e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2977,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.71875,
+      "step": 234
+    },
+    {
+      "epoch": 0.2883435582822086,
+      "grad_norm": 37.47355891059506,
+      "learning_rate": 7.305785123966941e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -320.0,
+      "loss": 0.284,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.0625,
+      "step": 235
+    },
+    {
+      "epoch": 0.28957055214723926,
+      "grad_norm": 53.75756263750908,
+      "learning_rate": 7.302479338842975e-07,
+      "logits/chosen": -0.0238037109375,
+      "logits/rejected": -0.2080078125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -368.0,
+      "loss": 0.4235,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.1875,
+      "step": 236
+    },
+    {
+      "epoch": 0.2907975460122699,
+      "grad_norm": 34.35888091553625,
+      "learning_rate": 7.299173553719008e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2606,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.09375,
+      "step": 237
+    },
+    {
+      "epoch": 0.2920245398773006,
+      "grad_norm": 42.88801905010467,
+      "learning_rate": 7.295867768595041e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3538,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7578125,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -3.90625,
+      "step": 238
+    },
+    {
+      "epoch": 0.29325153374233126,
+      "grad_norm": 42.29179859412283,
+      "learning_rate": 7.292561983471074e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3291,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.15625,
+      "step": 239
+    },
+    {
+      "epoch": 0.294478527607362,
+      "grad_norm": 37.481940316076816,
+      "learning_rate": 7.289256198347107e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -360.0,
+      "loss": 0.2328,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.890625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -4.6875,
+      "step": 240
+    },
+    {
+      "epoch": 0.29570552147239265,
+      "grad_norm": 37.13709897048156,
+      "learning_rate": 7.28595041322314e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2698,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.25,
+      "step": 241
+    },
+    {
+      "epoch": 0.2969325153374233,
+      "grad_norm": 30.985673644793227,
+      "learning_rate": 7.282644628099174e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -318.0,
+      "loss": 0.228,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.84375,
+      "step": 242
+    },
+    {
+      "epoch": 0.298159509202454,
+      "grad_norm": 53.9279211325065,
+      "learning_rate": 7.279338842975206e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -364.0,
+      "loss": 0.3208,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.34375,
+      "step": 243
+    },
+    {
+      "epoch": 0.29938650306748466,
+      "grad_norm": 44.157193824232536,
+      "learning_rate": 7.276033057851239e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -346.0,
+      "loss": 0.4473,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.125,
+      "step": 244
+    },
+    {
+      "epoch": 0.3006134969325153,
+      "grad_norm": 36.965116220324916,
+      "learning_rate": 7.272727272727272e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2506,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.46875,
+      "step": 245
+    },
+    {
+      "epoch": 0.301840490797546,
+      "grad_norm": 35.79905789258026,
+      "learning_rate": 7.269421487603305e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.292,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.078125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.40625,
+      "step": 246
+    },
+    {
+      "epoch": 0.3030674846625767,
+      "grad_norm": 46.030369589419166,
+      "learning_rate": 7.266115702479338e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3237,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.75,
+      "step": 247
+    },
+    {
+      "epoch": 0.3042944785276074,
+      "grad_norm": 38.420936945371935,
+      "learning_rate": 7.262809917355372e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2497,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.28125,
+      "step": 248
+    },
+    {
+      "epoch": 0.30552147239263805,
+      "grad_norm": 42.67532255005307,
+      "learning_rate": 7.259504132231405e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -476.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2862,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.328125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.3125,
+      "step": 249
+    },
+    {
+      "epoch": 0.3067484662576687,
+      "grad_norm": 40.02525959257945,
+      "learning_rate": 7.256198347107438e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -342.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3299,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -4.40625,
+      "step": 250
+    },
+    {
+      "epoch": 0.3079754601226994,
+      "grad_norm": 43.266430285019,
+      "learning_rate": 7.252892561983471e-07,
+      "logits/chosen": -0.041015625,
+      "logits/rejected": -0.2138671875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3567,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.75,
+      "step": 251
+    },
+    {
+      "epoch": 0.30920245398773005,
+      "grad_norm": 37.89906733852128,
+      "learning_rate": 7.249586776859503e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2982,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.7734375,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -3.859375,
+      "step": 252
+    },
+    {
+      "epoch": 0.3104294478527607,
+      "grad_norm": 33.28018552295872,
+      "learning_rate": 7.246280991735536e-07,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2522,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.46875,
+      "step": 253
+    },
+    {
+      "epoch": 0.3116564417177914,
+      "grad_norm": 42.254563825083,
+      "learning_rate": 7.24297520661157e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3681,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.15625,
+      "step": 254
+    },
+    {
+      "epoch": 0.3128834355828221,
+      "grad_norm": 39.895653943663504,
+      "learning_rate": 7.239669421487603e-07,
+      "logits/chosen": -0.1396484375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3481,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.8359375,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.46875,
+      "step": 255
+    },
+    {
+      "epoch": 0.3141104294478528,
+      "grad_norm": 37.11298367671905,
+      "learning_rate": 7.236363636363636e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2374,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.34375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.0625,
+      "step": 256
+    },
+    {
+      "epoch": 0.31533742331288345,
+      "grad_norm": 44.74542880830349,
+      "learning_rate": 7.233057851239669e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -426.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3118,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -3.546875,
+      "step": 257
+    },
+    {
+      "epoch": 0.3165644171779141,
+      "grad_norm": 34.05504946265699,
+      "learning_rate": 7.229752066115702e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2467,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.8984375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -3.859375,
+      "step": 258
+    },
+    {
+      "epoch": 0.3177914110429448,
+      "grad_norm": 35.06646898219457,
+      "learning_rate": 7.226446280991736e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -344.0,
+      "loss": 0.278,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.96875,
+      "step": 259
+    },
+    {
+      "epoch": 0.31901840490797545,
+      "grad_norm": 48.15030024240453,
+      "learning_rate": 7.223140495867769e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3099,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -3.640625,
+      "step": 260
+    },
+    {
+      "epoch": 0.3202453987730061,
+      "grad_norm": 43.183782239877885,
+      "learning_rate": 7.219834710743801e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3207,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.71875,
+      "step": 261
+    },
+    {
+      "epoch": 0.3214723926380368,
+      "grad_norm": 43.81473087786423,
+      "learning_rate": 7.216528925619834e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3077,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -3.484375,
+      "step": 262
+    },
+    {
+      "epoch": 0.3226993865030675,
+      "grad_norm": 45.763039013247074,
+      "learning_rate": 7.213223140495867e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -304.0,
+      "loss": 0.345,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -3.265625,
+      "step": 263
+    },
+    {
+      "epoch": 0.3239263803680982,
+      "grad_norm": 41.54807658858699,
+      "learning_rate": 7.2099173553719e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -288.0,
+      "loss": 0.3193,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -3.234375,
+      "step": 264
+    },
+    {
+      "epoch": 0.32515337423312884,
+      "grad_norm": 43.72770395576322,
+      "learning_rate": 7.206611570247934e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -266.0,
+      "loss": 0.296,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 4.96875,
+      "rewards/rejected": -3.515625,
+      "step": 265
+    },
+    {
+      "epoch": 0.3263803680981595,
+      "grad_norm": 49.68472762312363,
+      "learning_rate": 7.203305785123967e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -316.0,
+      "logps/rejected": -292.0,
+      "loss": 0.41,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 0.68359375,
+      "rewards/margins": 3.875,
+      "rewards/rejected": -3.1875,
+      "step": 266
+    },
+    {
+      "epoch": 0.3276073619631902,
+      "grad_norm": 54.21686609921306,
+      "learning_rate": 7.2e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3765,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.828125,
+      "rewards/margins": 4.8125,
+      "rewards/rejected": -2.984375,
+      "step": 267
+    },
+    {
+      "epoch": 0.32883435582822085,
+      "grad_norm": 36.013473849533085,
+      "learning_rate": 7.196694214876033e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2353,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.328125,
+      "step": 268
+    },
+    {
+      "epoch": 0.3300613496932515,
+      "grad_norm": 37.07985683341205,
+      "learning_rate": 7.193388429752066e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -400.0,
+      "logps/rejected": -292.0,
+      "loss": 0.229,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -3.46875,
+      "step": 269
+    },
+    {
+      "epoch": 0.3312883435582822,
+      "grad_norm": 43.39229519288019,
+      "learning_rate": 7.190082644628098e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -326.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3693,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 4.1875,
+      "rewards/rejected": -2.875,
+      "step": 270
+    },
+    {
+      "epoch": 0.3325153374233129,
+      "grad_norm": 33.45853348793219,
+      "learning_rate": 7.186776859504132e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2511,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.34375,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.234375,
+      "step": 271
+    },
+    {
+      "epoch": 0.3337423312883436,
+      "grad_norm": 36.86544969536763,
+      "learning_rate": 7.183471074380165e-07,
+      "logits/chosen": -0.1025390625,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2449,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.640625,
+      "step": 272
+    },
+    {
+      "epoch": 0.33496932515337424,
+      "grad_norm": 45.60464590137911,
+      "learning_rate": 7.180165289256198e-07,
+      "logits/chosen": -0.05029296875,
+      "logits/rejected": -0.1357421875,
+      "logps/chosen": -334.0,
+      "logps/rejected": -292.0,
+      "loss": 0.4336,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.4375,
+      "step": 273
+    },
+    {
+      "epoch": 0.3361963190184049,
+      "grad_norm": 38.72003382453235,
+      "learning_rate": 7.176859504132231e-07,
+      "logits/chosen": -0.06591796875,
+      "logits/rejected": -0.21875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3223,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.359375,
+      "step": 274
+    },
+    {
+      "epoch": 0.3374233128834356,
+      "grad_norm": 41.46753745108232,
+      "learning_rate": 7.173553719008264e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -316.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3348,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.5,
+      "step": 275
+    },
+    {
+      "epoch": 0.33865030674846625,
+      "grad_norm": 29.942801211334718,
+      "learning_rate": 7.170247933884297e-07,
+      "logits/chosen": -0.056396484375,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2861,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.84375,
+      "step": 276
+    },
+    {
+      "epoch": 0.3398773006134969,
+      "grad_norm": 36.649252691867474,
+      "learning_rate": 7.166942148760331e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2238,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.3125,
+      "step": 277
+    },
+    {
+      "epoch": 0.3411042944785276,
+      "grad_norm": 41.64219712301306,
+      "learning_rate": 7.163636363636364e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2881,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.125,
+      "step": 278
+    },
+    {
+      "epoch": 0.3423312883435583,
+      "grad_norm": 25.028100153022006,
+      "learning_rate": 7.160330578512396e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1819,
+      "rewards/accuracies": 0.9453125,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.875,
+      "step": 279
+    },
+    {
+      "epoch": 0.34355828220858897,
+      "grad_norm": 43.34524582256192,
+      "learning_rate": 7.157024793388429e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3811,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.625,
+      "step": 280
+    },
+    {
+      "epoch": 0.34478527607361964,
+      "grad_norm": 39.57499610045821,
+      "learning_rate": 7.153719008264462e-07,
+      "logits/chosen": -0.12158203125,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3441,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.9375,
+      "step": 281
+    },
+    {
+      "epoch": 0.3460122699386503,
+      "grad_norm": 32.558914994483004,
+      "learning_rate": 7.150413223140495e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -340.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2162,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.0625,
+      "step": 282
+    },
+    {
+      "epoch": 0.347239263803681,
+      "grad_norm": 33.09320412279389,
+      "learning_rate": 7.147107438016529e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2537,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.984375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.5,
+      "step": 283
+    },
+    {
+      "epoch": 0.34846625766871164,
+      "grad_norm": 40.734540729970675,
+      "learning_rate": 7.143801652892562e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -310.0,
+      "loss": 0.272,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.84375,
+      "step": 284
+    },
+    {
+      "epoch": 0.3496932515337423,
+      "grad_norm": 30.9410258264426,
+      "learning_rate": 7.140495867768595e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2413,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.96875,
+      "step": 285
+    },
+    {
+      "epoch": 0.350920245398773,
+      "grad_norm": 37.246332837421065,
+      "learning_rate": 7.137190082644628e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3337,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.09375,
+      "step": 286
+    },
+    {
+      "epoch": 0.3521472392638037,
+      "grad_norm": 40.87355956326826,
+      "learning_rate": 7.133884297520661e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2981,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.68359375,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.9375,
+      "step": 287
+    },
+    {
+      "epoch": 0.35337423312883437,
+      "grad_norm": 37.148615900618886,
+      "learning_rate": 7.130578512396693e-07,
+      "logits/chosen": -0.05908203125,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3085,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -4.59375,
+      "step": 288
+    },
+    {
+      "epoch": 0.35460122699386504,
+      "grad_norm": 29.32351813266847,
+      "learning_rate": 7.127272727272727e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -316.0,
+      "loss": 0.1987,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.34375,
+      "step": 289
+    },
+    {
+      "epoch": 0.3558282208588957,
+      "grad_norm": 34.03478605235303,
+      "learning_rate": 7.12396694214876e-07,
+      "logits/chosen": -0.359375,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2024,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.71875,
+      "step": 290
+    },
+    {
+      "epoch": 0.3570552147239264,
+      "grad_norm": 38.57928856233812,
+      "learning_rate": 7.120661157024793e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3224,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -4.5,
+      "step": 291
+    },
+    {
+      "epoch": 0.35828220858895704,
+      "grad_norm": 36.5133266014805,
+      "learning_rate": 7.117355371900826e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2451,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.03125,
+      "step": 292
+    },
+    {
+      "epoch": 0.3595092024539877,
+      "grad_norm": 31.073871401773395,
+      "learning_rate": 7.114049586776859e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2162,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.25,
+      "step": 293
+    },
+    {
+      "epoch": 0.36073619631901843,
+      "grad_norm": 31.419020907335653,
+      "learning_rate": 7.110743801652892e-07,
+      "logits/chosen": -0.427734375,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -464.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2389,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.3828125,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.53125,
+      "step": 294
+    },
+    {
+      "epoch": 0.3619631901840491,
+      "grad_norm": 33.66575248055817,
+      "learning_rate": 7.107438016528926e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2242,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.03125,
+      "step": 295
+    },
+    {
+      "epoch": 0.36319018404907977,
+      "grad_norm": 40.06405186289625,
+      "learning_rate": 7.104132231404959e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3209,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.1875,
+      "step": 296
+    },
+    {
+      "epoch": 0.36441717791411044,
+      "grad_norm": 30.371449913820346,
+      "learning_rate": 7.100826446280991e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -352.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2354,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.5,
+      "step": 297
+    },
+    {
+      "epoch": 0.3656441717791411,
+      "grad_norm": 40.780831154790015,
+      "learning_rate": 7.097520661157024e-07,
+      "logits/chosen": -0.126953125,
+      "logits/rejected": -0.23046875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2901,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.890625,
+      "step": 298
+    },
+    {
+      "epoch": 0.36687116564417177,
+      "grad_norm": 48.39132901882327,
+      "learning_rate": 7.094214876033057e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3427,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -4.0625,
+      "step": 299
+    },
+    {
+      "epoch": 0.36809815950920244,
+      "grad_norm": 39.25730980280545,
+      "learning_rate": 7.09090909090909e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3287,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.21875,
+      "step": 300
+    },
+    {
+      "epoch": 0.3693251533742331,
+      "grad_norm": 34.87785766178715,
+      "learning_rate": 7.087603305785124e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -448.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2823,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.1875,
+      "step": 301
+    },
+    {
+      "epoch": 0.37055214723926383,
+      "grad_norm": 30.464291800676158,
+      "learning_rate": 7.084297520661157e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2182,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -4.15625,
+      "step": 302
+    },
+    {
+      "epoch": 0.3717791411042945,
+      "grad_norm": 45.78814856158198,
+      "learning_rate": 7.08099173553719e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3512,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.40625,
+      "step": 303
+    },
+    {
+      "epoch": 0.37300613496932516,
+      "grad_norm": 33.26834608933062,
+      "learning_rate": 7.077685950413223e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2448,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.53125,
+      "step": 304
+    },
+    {
+      "epoch": 0.37423312883435583,
+      "grad_norm": 35.609604648743236,
+      "learning_rate": 7.074380165289256e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2744,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.625,
+      "step": 305
+    },
+    {
+      "epoch": 0.3754601226993865,
+      "grad_norm": 43.837358024220855,
+      "learning_rate": 7.071074380165288e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2737,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.5,
+      "step": 306
+    },
+    {
+      "epoch": 0.37668711656441717,
+      "grad_norm": 39.17650253748135,
+      "learning_rate": 7.067768595041322e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3713,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.71875,
+      "step": 307
+    },
+    {
+      "epoch": 0.37791411042944784,
+      "grad_norm": 36.807676017880326,
+      "learning_rate": 7.064462809917355e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3165,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.94140625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -4.0625,
+      "step": 308
+    },
+    {
+      "epoch": 0.3791411042944785,
+      "grad_norm": 32.288551419968854,
+      "learning_rate": 7.061157024793388e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2283,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.6484375,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -4.90625,
+      "step": 309
+    },
+    {
+      "epoch": 0.3803680981595092,
+      "grad_norm": 47.130716517812814,
+      "learning_rate": 7.057851239669421e-07,
+      "logits/chosen": -0.0908203125,
+      "logits/rejected": -0.1884765625,
+      "logps/chosen": -312.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3596,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.15234375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.6875,
+      "step": 310
+    },
+    {
+      "epoch": 0.3815950920245399,
+      "grad_norm": 34.12148114082454,
+      "learning_rate": 7.054545454545454e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2543,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -5.375,
+      "step": 311
+    },
+    {
+      "epoch": 0.38282208588957056,
+      "grad_norm": 44.35564676375722,
+      "learning_rate": 7.051239669421488e-07,
+      "logits/chosen": -0.125,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -362.0,
+      "loss": 0.4044,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -4.5,
+      "step": 312
+    },
+    {
+      "epoch": 0.38404907975460123,
+      "grad_norm": 33.82923751151301,
+      "learning_rate": 7.047933884297521e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2685,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.8828125,
+      "rewards/margins": 7.28125,
+      "rewards/rejected": -5.375,
+      "step": 313
+    },
+    {
+      "epoch": 0.3852760736196319,
+      "grad_norm": 33.17604766084792,
+      "learning_rate": 7.044628099173554e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -456.0,
+      "logps/rejected": -344.0,
+      "loss": 0.1847,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 2.25,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -5.71875,
+      "step": 314
+    },
+    {
+      "epoch": 0.38650306748466257,
+      "grad_norm": 38.90508166705689,
+      "learning_rate": 7.041322314049586e-07,
+      "logits/chosen": -0.06396484375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -286.0,
+      "loss": 0.2412,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.5546875,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.21875,
+      "step": 315
+    },
+    {
+      "epoch": 0.38773006134969323,
+      "grad_norm": 31.950078808944912,
+      "learning_rate": 7.038016528925619e-07,
+      "logits/chosen": -0.06982421875,
+      "logits/rejected": -0.2158203125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2551,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.4375,
+      "step": 316
+    },
+    {
+      "epoch": 0.3889570552147239,
+      "grad_norm": 34.528377745666106,
+      "learning_rate": 7.034710743801652e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2296,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.875,
+      "step": 317
+    },
+    {
+      "epoch": 0.3901840490797546,
+      "grad_norm": 46.9506582652498,
+      "learning_rate": 7.031404958677686e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3503,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.71875,
+      "step": 318
+    },
+    {
+      "epoch": 0.3914110429447853,
+      "grad_norm": 26.867054074267944,
+      "learning_rate": 7.028099173553719e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1701,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 7.4375,
+      "rewards/rejected": -6.125,
+      "step": 319
+    },
+    {
+      "epoch": 0.39263803680981596,
+      "grad_norm": 40.540053142562215,
+      "learning_rate": 7.024793388429752e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -318.0,
+      "logps/rejected": -274.0,
+      "loss": 0.3046,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -5.3125,
+      "step": 320
+    },
+    {
+      "epoch": 0.39386503067484663,
+      "grad_norm": 46.49642012587718,
+      "learning_rate": 7.021487603305785e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -384.0,
+      "loss": 0.2961,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.96875,
+      "step": 321
+    },
+    {
+      "epoch": 0.3950920245398773,
+      "grad_norm": 33.72622281793196,
+      "learning_rate": 7.018181818181818e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2496,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.3125,
+      "step": 322
+    },
+    {
+      "epoch": 0.39631901840490796,
+      "grad_norm": 39.04984997032008,
+      "learning_rate": 7.014876033057851e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3837,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -4.3125,
+      "step": 323
+    },
+    {
+      "epoch": 0.39754601226993863,
+      "grad_norm": 33.57036215026917,
+      "learning_rate": 7.011570247933884e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2714,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.5,
+      "step": 324
+    },
+    {
+      "epoch": 0.3987730061349693,
+      "grad_norm": 45.64995516248102,
+      "learning_rate": 7.008264462809917e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.2177734375,
+      "logps/chosen": -328.0,
+      "logps/rejected": -290.0,
+      "loss": 0.4112,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.6875,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -4.53125,
+      "step": 325
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 34.28943945713478,
+      "learning_rate": 7.00495867768595e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2004,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.09375,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -5.03125,
+      "step": 326
+    },
+    {
+      "epoch": 0.4012269938650307,
+      "grad_norm": 48.50197304580951,
+      "learning_rate": 7.001652892561983e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3949,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.6875,
+      "step": 327
+    },
+    {
+      "epoch": 0.40245398773006136,
+      "grad_norm": 40.72777385232644,
+      "learning_rate": 6.998347107438016e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -272.0,
+      "loss": 0.3146,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -4.3125,
+      "step": 328
+    },
+    {
+      "epoch": 0.403680981595092,
+      "grad_norm": 35.55525281765151,
+      "learning_rate": 6.995041322314049e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -454.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2936,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -4.90625,
+      "step": 329
+    },
+    {
+      "epoch": 0.4049079754601227,
+      "grad_norm": 42.0221302571432,
+      "learning_rate": 6.991735537190083e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2435,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9765625,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -4.71875,
+      "step": 330
+    },
+    {
+      "epoch": 0.40613496932515336,
+      "grad_norm": 32.84700653520424,
+      "learning_rate": 6.988429752066116e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2501,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.9375,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -4.375,
+      "step": 331
+    },
+    {
+      "epoch": 0.40736196319018403,
+      "grad_norm": 27.379927529722597,
+      "learning_rate": 6.985123966942148e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2017,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.46875,
+      "step": 332
+    },
+    {
+      "epoch": 0.4085889570552147,
+      "grad_norm": 36.70659629756495,
+      "learning_rate": 6.981818181818181e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3078,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -3.84375,
+      "step": 333
+    },
+    {
+      "epoch": 0.4098159509202454,
+      "grad_norm": 34.15971280742034,
+      "learning_rate": 6.978512396694214e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -338.0,
+      "loss": 0.273,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -4.5,
+      "step": 334
+    },
+    {
+      "epoch": 0.4110429447852761,
+      "grad_norm": 37.2375068536928,
+      "learning_rate": 6.975206611570247e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2248,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.03125,
+      "step": 335
+    },
+    {
+      "epoch": 0.41226993865030676,
+      "grad_norm": 36.84981287485191,
+      "learning_rate": 6.971900826446281e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2764,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.21875,
+      "step": 336
+    },
+    {
+      "epoch": 0.4134969325153374,
+      "grad_norm": 46.94122451213228,
+      "learning_rate": 6.968595041322314e-07,
+      "logits/chosen": -0.09375,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -344.0,
+      "logps/rejected": -322.0,
+      "loss": 0.389,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.921875,
+      "step": 337
+    },
+    {
+      "epoch": 0.4147239263803681,
+      "grad_norm": 38.247859512270246,
+      "learning_rate": 6.965289256198347e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3295,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.0,
+      "step": 338
+    },
+    {
+      "epoch": 0.41595092024539876,
+      "grad_norm": 40.84741702132196,
+      "learning_rate": 6.96198347107438e-07,
+      "logits/chosen": -0.07470703125,
+      "logits/rejected": -0.2109375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3275,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.15625,
+      "step": 339
+    },
+    {
+      "epoch": 0.4171779141104294,
+      "grad_norm": 43.700613878566976,
+      "learning_rate": 6.958677685950413e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2879,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.09375,
+      "step": 340
+    },
+    {
+      "epoch": 0.41840490797546015,
+      "grad_norm": 31.54531879054049,
+      "learning_rate": 6.955371900826445e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2805,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.65625,
+      "step": 341
+    },
+    {
+      "epoch": 0.4196319018404908,
+      "grad_norm": 39.0902727508844,
+      "learning_rate": 6.952066115702479e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -328.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3322,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.1875,
+      "step": 342
+    },
+    {
+      "epoch": 0.4208588957055215,
+      "grad_norm": 39.32051666516945,
+      "learning_rate": 6.948760330578512e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3545,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.921875,
+      "step": 343
+    },
+    {
+      "epoch": 0.42208588957055215,
+      "grad_norm": 42.12526028461151,
+      "learning_rate": 6.945454545454545e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3584,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.0625,
+      "step": 344
+    },
+    {
+      "epoch": 0.4233128834355828,
+      "grad_norm": 39.02371998906191,
+      "learning_rate": 6.942148760330578e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3288,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.3125,
+      "step": 345
+    },
+    {
+      "epoch": 0.4245398773006135,
+      "grad_norm": 41.19695447421548,
+      "learning_rate": 6.938842975206611e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -332.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3569,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -4.125,
+      "step": 346
+    },
+    {
+      "epoch": 0.42576687116564416,
+      "grad_norm": 28.472619739287644,
+      "learning_rate": 6.935537190082645e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -360.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2332,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.25,
+      "step": 347
+    },
+    {
+      "epoch": 0.4269938650306748,
+      "grad_norm": 33.19419614917591,
+      "learning_rate": 6.932231404958678e-07,
+      "logits/chosen": -0.24609375,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2594,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.91796875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.59375,
+      "step": 348
+    },
+    {
+      "epoch": 0.42822085889570555,
+      "grad_norm": 38.111059722172605,
+      "learning_rate": 6.928925619834711e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3255,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.34375,
+      "step": 349
+    },
+    {
+      "epoch": 0.4294478527607362,
+      "grad_norm": 37.49784569992658,
+      "learning_rate": 6.925619834710743e-07,
+      "logits/chosen": -0.0986328125,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2857,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.28125,
+      "step": 350
+    },
+    {
+      "epoch": 0.4306748466257669,
+      "grad_norm": 42.647163206760126,
+      "learning_rate": 6.922314049586776e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -336.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3505,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.5078125,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -4.78125,
+      "step": 351
+    },
+    {
+      "epoch": 0.43190184049079755,
+      "grad_norm": 36.26120743659289,
+      "learning_rate": 6.919008264462809e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3453,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.03125,
+      "step": 352
+    },
+    {
+      "epoch": 0.4331288343558282,
+      "grad_norm": 30.933441803520967,
+      "learning_rate": 6.915702479338843e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2406,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.96875,
+      "step": 353
+    },
+    {
+      "epoch": 0.4343558282208589,
+      "grad_norm": 34.874499707263155,
+      "learning_rate": 6.912396694214876e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2963,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.03125,
+      "step": 354
+    },
+    {
+      "epoch": 0.43558282208588955,
+      "grad_norm": 39.848840136754845,
+      "learning_rate": 6.909090909090909e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -340.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3581,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.53125,
+      "step": 355
+    },
+    {
+      "epoch": 0.4368098159509202,
+      "grad_norm": 44.03875893573232,
+      "learning_rate": 6.905785123966942e-07,
+      "logits/chosen": -0.0966796875,
+      "logits/rejected": -0.240234375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3406,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.03125,
+      "step": 356
+    },
+    {
+      "epoch": 0.43803680981595094,
+      "grad_norm": 36.68968799262988,
+      "learning_rate": 6.902479338842975e-07,
+      "logits/chosen": -0.3515625,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2926,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -4.9375,
+      "step": 357
+    },
+    {
+      "epoch": 0.4392638036809816,
+      "grad_norm": 36.204158508103596,
+      "learning_rate": 6.899173553719008e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2673,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.40625,
+      "step": 358
+    },
+    {
+      "epoch": 0.4404907975460123,
+      "grad_norm": 40.16138545949422,
+      "learning_rate": 6.89586776859504e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.953125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.09375,
+      "step": 359
+    },
+    {
+      "epoch": 0.44171779141104295,
+      "grad_norm": 35.21609251200673,
+      "learning_rate": 6.892561983471074e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2614,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.9375,
+      "step": 360
+    },
+    {
+      "epoch": 0.4429447852760736,
+      "grad_norm": 33.95190421090596,
+      "learning_rate": 6.889256198347107e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2599,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -4.8125,
+      "step": 361
+    },
+    {
+      "epoch": 0.4441717791411043,
+      "grad_norm": 32.41395226001234,
+      "learning_rate": 6.88595041322314e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2625,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -4.75,
+      "step": 362
+    },
+    {
+      "epoch": 0.44539877300613495,
+      "grad_norm": 31.06408733284358,
+      "learning_rate": 6.882644628099173e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2362,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.375,
+      "step": 363
+    },
+    {
+      "epoch": 0.4466257668711656,
+      "grad_norm": 40.778000027843554,
+      "learning_rate": 6.879338842975206e-07,
+      "logits/chosen": -0.07470703125,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2885,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.28125,
+      "step": 364
+    },
+    {
+      "epoch": 0.44785276073619634,
+      "grad_norm": 30.8545654380382,
+      "learning_rate": 6.87603305785124e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -298.0,
+      "loss": 0.2419,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.53125,
+      "step": 365
+    },
+    {
+      "epoch": 0.449079754601227,
+      "grad_norm": 35.25923514975111,
+      "learning_rate": 6.872727272727273e-07,
+      "logits/chosen": -0.11376953125,
+      "logits/rejected": -0.21484375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3324,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.71875,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.703125,
+      "step": 366
+    },
+    {
+      "epoch": 0.4503067484662577,
+      "grad_norm": 29.75727838549342,
+      "learning_rate": 6.869421487603306e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2265,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.25,
+      "step": 367
+    },
+    {
+      "epoch": 0.45153374233128835,
+      "grad_norm": 41.34860966923763,
+      "learning_rate": 6.866115702479338e-07,
+      "logits/chosen": -0.060546875,
+      "logits/rejected": -0.11328125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3251,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.1875,
+      "step": 368
+    },
+    {
+      "epoch": 0.452760736196319,
+      "grad_norm": 39.802701196225414,
+      "learning_rate": 6.862809917355371e-07,
+      "logits/chosen": -0.08349609375,
+      "logits/rejected": -0.19140625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3103,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -4.375,
+      "step": 369
+    },
+    {
+      "epoch": 0.4539877300613497,
+      "grad_norm": 35.44626532768664,
+      "learning_rate": 6.859504132231404e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3001,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.0,
+      "step": 370
+    },
+    {
+      "epoch": 0.45521472392638035,
+      "grad_norm": 34.081640470685045,
+      "learning_rate": 6.856198347107438e-07,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -458.0,
+      "logps/rejected": -356.0,
+      "loss": 0.268,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.734375,
+      "step": 371
+    },
+    {
+      "epoch": 0.456441717791411,
+      "grad_norm": 32.10146522336104,
+      "learning_rate": 6.852892561983471e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2227,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.15625,
+      "step": 372
+    },
+    {
+      "epoch": 0.45766871165644174,
+      "grad_norm": 38.40212256087746,
+      "learning_rate": 6.849586776859504e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -360.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3491,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.88671875,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -4.40625,
+      "step": 373
+    },
+    {
+      "epoch": 0.4588957055214724,
+      "grad_norm": 33.89078532536025,
+      "learning_rate": 6.846280991735537e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2439,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.25,
+      "step": 374
+    },
+    {
+      "epoch": 0.4601226993865031,
+      "grad_norm": 38.36603856770271,
+      "learning_rate": 6.84297520661157e-07,
+      "logits/chosen": -0.24609375,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2743,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.78125,
+      "step": 375
+    },
+    {
+      "epoch": 0.46134969325153374,
+      "grad_norm": 32.73897251445023,
+      "learning_rate": 6.839669421487603e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -342.0,
+      "loss": 0.231,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -3.9375,
+      "step": 376
+    },
+    {
+      "epoch": 0.4625766871165644,
+      "grad_norm": 36.44881761827906,
+      "learning_rate": 6.836363636363636e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -350.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2817,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.8125,
+      "step": 377
+    },
+    {
+      "epoch": 0.4638036809815951,
+      "grad_norm": 36.92866834323212,
+      "learning_rate": 6.833057851239669e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3001,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.5,
+      "step": 378
+    },
+    {
+      "epoch": 0.46503067484662575,
+      "grad_norm": 32.73144242036404,
+      "learning_rate": 6.829752066115702e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2471,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7578125,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.0625,
+      "step": 379
+    },
+    {
+      "epoch": 0.4662576687116564,
+      "grad_norm": 34.6763093414022,
+      "learning_rate": 6.826446280991735e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2659,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -4.96875,
+      "step": 380
+    },
+    {
+      "epoch": 0.46748466257668714,
+      "grad_norm": 34.98446202126656,
+      "learning_rate": 6.823140495867768e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3193,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.59375,
+      "step": 381
+    },
+    {
+      "epoch": 0.4687116564417178,
+      "grad_norm": 33.068661800772716,
+      "learning_rate": 6.819834710743801e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.40625,
+      "step": 382
+    },
+    {
+      "epoch": 0.4699386503067485,
+      "grad_norm": 37.187464031340724,
+      "learning_rate": 6.816528925619835e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -378.0,
+      "loss": 0.33,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.625,
+      "step": 383
+    },
+    {
+      "epoch": 0.47116564417177914,
+      "grad_norm": 35.133860722289874,
+      "learning_rate": 6.813223140495868e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2466,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -6.5,
+      "step": 384
+    },
+    {
+      "epoch": 0.4723926380368098,
+      "grad_norm": 40.48440514462701,
+      "learning_rate": 6.809917355371901e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3252,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.2578125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -5.78125,
+      "step": 385
+    },
+    {
+      "epoch": 0.4736196319018405,
+      "grad_norm": 40.628340663991764,
+      "learning_rate": 6.806611570247933e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3025,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -0.359375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -6.0625,
+      "step": 386
+    },
+    {
+      "epoch": 0.47484662576687114,
+      "grad_norm": 51.72741620856106,
+      "learning_rate": 6.803305785123966e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.90625,
+      "step": 387
+    },
+    {
+      "epoch": 0.47607361963190187,
+      "grad_norm": 43.50814776279542,
+      "learning_rate": 6.8e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3977,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.05029296875,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -5.09375,
+      "step": 388
+    },
+    {
+      "epoch": 0.47730061349693254,
+      "grad_norm": 44.9083628878753,
+      "learning_rate": 6.796694214876033e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3913,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.44140625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.625,
+      "step": 389
+    },
+    {
+      "epoch": 0.4785276073619632,
+      "grad_norm": 42.893730022875914,
+      "learning_rate": 6.793388429752066e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3759,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": -0.263671875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -6.0,
+      "step": 390
+    },
+    {
+      "epoch": 0.47975460122699387,
+      "grad_norm": 46.56902163562806,
+      "learning_rate": 6.790082644628099e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4122,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.3984375,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.5625,
+      "step": 391
+    },
+    {
+      "epoch": 0.48098159509202454,
+      "grad_norm": 32.46380915314015,
+      "learning_rate": 6.786776859504132e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2695,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.65625,
+      "step": 392
+    },
+    {
+      "epoch": 0.4822085889570552,
+      "grad_norm": 36.411647842869584,
+      "learning_rate": 6.783471074380165e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2936,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.66015625,
+      "rewards/margins": 6.125,
+      "rewards/rejected": -5.46875,
+      "step": 393
+    },
+    {
+      "epoch": 0.4834355828220859,
+      "grad_norm": 36.43693693488016,
+      "learning_rate": 6.780165289256199e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2504,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.7265625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.34375,
+      "step": 394
+    },
+    {
+      "epoch": 0.48466257668711654,
+      "grad_norm": 36.92302249692751,
+      "learning_rate": 6.776859504132231e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3275,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.96875,
+      "step": 395
+    },
+    {
+      "epoch": 0.48588957055214727,
+      "grad_norm": 31.266490118113346,
+      "learning_rate": 6.773553719008264e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -338.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2215,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.462890625,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.6875,
+      "step": 396
+    },
+    {
+      "epoch": 0.48711656441717793,
+      "grad_norm": 34.937182799458625,
+      "learning_rate": 6.770247933884297e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2752,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.8125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.65625,
+      "step": 397
+    },
+    {
+      "epoch": 0.4883435582822086,
+      "grad_norm": 38.419527775162116,
+      "learning_rate": 6.76694214876033e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3231,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.0625,
+      "step": 398
+    },
+    {
+      "epoch": 0.48957055214723927,
+      "grad_norm": 35.54458075122045,
+      "learning_rate": 6.763636363636363e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -330.0,
+      "loss": 0.306,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.9375,
+      "step": 399
+    },
+    {
+      "epoch": 0.49079754601226994,
+      "grad_norm": 47.751486198572856,
+      "learning_rate": 6.760330578512397e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3896,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.90234375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.46875,
+      "step": 400
+    },
+    {
+      "epoch": 0.4920245398773006,
+      "grad_norm": 31.519656830052252,
+      "learning_rate": 6.75702479338843e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -444.0,
+      "logps/rejected": -372.0,
+      "loss": 0.2337,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.34375,
+      "step": 401
+    },
+    {
+      "epoch": 0.49325153374233127,
+      "grad_norm": 36.11809592072752,
+      "learning_rate": 6.753719008264463e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3095,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.78125,
+      "step": 402
+    },
+    {
+      "epoch": 0.49447852760736194,
+      "grad_norm": 38.21909292164752,
+      "learning_rate": 6.750413223140496e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3169,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.875,
+      "step": 403
+    },
+    {
+      "epoch": 0.49570552147239266,
+      "grad_norm": 32.22630836566114,
+      "learning_rate": 6.747107438016528e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2283,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.0,
+      "step": 404
+    },
+    {
+      "epoch": 0.49693251533742333,
+      "grad_norm": 37.593793152840426,
+      "learning_rate": 6.743801652892561e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3305,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.65625,
+      "step": 405
+    },
+    {
+      "epoch": 0.498159509202454,
+      "grad_norm": 41.53333095908596,
+      "learning_rate": 6.740495867768595e-07,
+      "logits/chosen": -0.115234375,
+      "logits/rejected": -0.2109375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3932,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.3125,
+      "step": 406
+    },
+    {
+      "epoch": 0.49938650306748467,
+      "grad_norm": 42.42876322706536,
+      "learning_rate": 6.737190082644628e-07,
+      "logits/chosen": -0.0986328125,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3513,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.140625,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -4.3125,
+      "step": 407
+    },
+    {
+      "epoch": 0.5006134969325153,
+      "grad_norm": 48.89324534838805,
+      "learning_rate": 6.733884297520661e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -348.0,
+      "loss": 0.4164,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 4.78125,
+      "rewards/rejected": -3.984375,
+      "step": 408
+    },
+    {
+      "epoch": 0.501840490797546,
+      "grad_norm": 45.10686444580896,
+      "learning_rate": 6.730578512396694e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3079,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.46875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -4.75,
+      "step": 409
+    },
+    {
+      "epoch": 0.5030674846625767,
+      "grad_norm": 38.62243714595701,
+      "learning_rate": 6.727272727272727e-07,
+      "logits/chosen": -0.119140625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2966,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.3125,
+      "step": 410
+    },
+    {
+      "epoch": 0.5042944785276073,
+      "grad_norm": 37.41136253841519,
+      "learning_rate": 6.72396694214876e-07,
+      "logits/chosen": -0.11083984375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -306.0,
+      "loss": 0.284,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.76953125,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.9375,
+      "step": 411
+    },
+    {
+      "epoch": 0.505521472392638,
+      "grad_norm": 47.69531455944101,
+      "learning_rate": 6.720661157024794e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3619,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -4.28125,
+      "step": 412
+    },
+    {
+      "epoch": 0.5067484662576687,
+      "grad_norm": 39.571364824526874,
+      "learning_rate": 6.717355371900826e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3069,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.59375,
+      "step": 413
+    },
+    {
+      "epoch": 0.5079754601226993,
+      "grad_norm": 34.32054957685153,
+      "learning_rate": 6.714049586776859e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2497,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.8125,
+      "step": 414
+    },
+    {
+      "epoch": 0.50920245398773,
+      "grad_norm": 35.2489180238823,
+      "learning_rate": 6.710743801652892e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2788,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.84375,
+      "step": 415
+    },
+    {
+      "epoch": 0.5104294478527608,
+      "grad_norm": 28.41248357906051,
+      "learning_rate": 6.707438016528925e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -310.0,
+      "loss": 0.186,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 0.8515625,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.75,
+      "step": 416
+    },
+    {
+      "epoch": 0.5116564417177915,
+      "grad_norm": 35.52909474934163,
+      "learning_rate": 6.704132231404958e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -434.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2995,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -5.53125,
+      "step": 417
+    },
+    {
+      "epoch": 0.5128834355828221,
+      "grad_norm": 36.29519060781132,
+      "learning_rate": 6.700826446280992e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -292.0,
+      "loss": 0.2828,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.09375,
+      "step": 418
+    },
+    {
+      "epoch": 0.5141104294478528,
+      "grad_norm": 31.87680208130441,
+      "learning_rate": 6.697520661157025e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2486,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.59375,
+      "step": 419
+    },
+    {
+      "epoch": 0.5153374233128835,
+      "grad_norm": 40.881697279170936,
+      "learning_rate": 6.694214876033058e-07,
+      "logits/chosen": -0.1162109375,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -344.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3244,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.75,
+      "step": 420
+    },
+    {
+      "epoch": 0.5165644171779141,
+      "grad_norm": 34.11082321167536,
+      "learning_rate": 6.69090909090909e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -312.0,
+      "loss": 0.339,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -4.8125,
+      "step": 421
+    },
+    {
+      "epoch": 0.5177914110429448,
+      "grad_norm": 35.18821029472038,
+      "learning_rate": 6.687603305785123e-07,
+      "logits/chosen": -0.0810546875,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3166,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.34375,
+      "step": 422
+    },
+    {
+      "epoch": 0.5190184049079755,
+      "grad_norm": 41.06855790218669,
+      "learning_rate": 6.684297520661156e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3606,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.96875,
+      "step": 423
+    },
+    {
+      "epoch": 0.5202453987730061,
+      "grad_norm": 36.724024878608596,
+      "learning_rate": 6.68099173553719e-07,
+      "logits/chosen": -0.123046875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3438,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.5625,
+      "step": 424
+    },
+    {
+      "epoch": 0.5214723926380368,
+      "grad_norm": 41.01181618114964,
+      "learning_rate": 6.677685950413223e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -342.0,
+      "loss": 0.4073,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.96875,
+      "step": 425
+    },
+    {
+      "epoch": 0.5226993865030675,
+      "grad_norm": 44.64404869891064,
+      "learning_rate": 6.674380165289256e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3035,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.46875,
+      "step": 426
+    },
+    {
+      "epoch": 0.5239263803680981,
+      "grad_norm": 91.8646628453638,
+      "learning_rate": 6.671074380165289e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -328.0,
+      "loss": 0.299,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.09375,
+      "step": 427
+    },
+    {
+      "epoch": 0.5251533742331288,
+      "grad_norm": 36.99313104379447,
+      "learning_rate": 6.667768595041322e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3507,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.66015625,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -5.21875,
+      "step": 428
+    },
+    {
+      "epoch": 0.5263803680981595,
+      "grad_norm": 33.0190569729192,
+      "learning_rate": 6.664462809917356e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -360.0,
+      "logps/rejected": -366.0,
+      "loss": 0.246,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.443359375,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -6.46875,
+      "step": 429
+    },
+    {
+      "epoch": 0.5276073619631901,
+      "grad_norm": 30.408983999620382,
+      "learning_rate": 6.661157024793388e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2248,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.9921875,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.84375,
+      "step": 430
+    },
+    {
+      "epoch": 0.5288343558282208,
+      "grad_norm": 47.87075432548269,
+      "learning_rate": 6.657851239669421e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -336.0,
+      "loss": 0.4545,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.45703125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -5.25,
+      "step": 431
+    },
+    {
+      "epoch": 0.5300613496932516,
+      "grad_norm": 43.65339869428465,
+      "learning_rate": 6.654545454545454e-07,
+      "logits/chosen": -0.0927734375,
+      "logits/rejected": -0.1435546875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3724,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.875,
+      "step": 432
+    },
+    {
+      "epoch": 0.5312883435582823,
+      "grad_norm": 45.133784704459096,
+      "learning_rate": 6.651239669421487e-07,
+      "logits/chosen": -0.078125,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3416,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.39453125,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.34375,
+      "step": 433
+    },
+    {
+      "epoch": 0.5325153374233129,
+      "grad_norm": 40.47930492236747,
+      "learning_rate": 6.64793388429752e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -346.0,
+      "loss": 0.254,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.53125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.25,
+      "step": 434
+    },
+    {
+      "epoch": 0.5337423312883436,
+      "grad_norm": 41.262979059220854,
+      "learning_rate": 6.644628099173554e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3119,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.40625,
+      "step": 435
+    },
+    {
+      "epoch": 0.5349693251533743,
+      "grad_norm": 38.67538014065472,
+      "learning_rate": 6.641322314049587e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2467,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 7.15625,
+      "rewards/rejected": -5.46875,
+      "step": 436
+    },
+    {
+      "epoch": 0.5361963190184049,
+      "grad_norm": 31.32861125197491,
+      "learning_rate": 6.63801652892562e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -338.0,
+      "loss": 0.248,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 6.65625,
+      "rewards/rejected": -5.5625,
+      "step": 437
+    },
+    {
+      "epoch": 0.5374233128834356,
+      "grad_norm": 42.810824724719026,
+      "learning_rate": 6.634710743801653e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.2392578125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3969,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.375,
+      "step": 438
+    },
+    {
+      "epoch": 0.5386503067484663,
+      "grad_norm": 36.79651022084742,
+      "learning_rate": 6.631404958677685e-07,
+      "logits/chosen": -0.083984375,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.1875,
+      "step": 439
+    },
+    {
+      "epoch": 0.5398773006134969,
+      "grad_norm": 49.84881532208057,
+      "learning_rate": 6.628099173553718e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.2333984375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3608,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.384765625,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -4.6875,
+      "step": 440
+    },
+    {
+      "epoch": 0.5411042944785276,
+      "grad_norm": 37.976634967176224,
+      "learning_rate": 6.624793388429752e-07,
+      "logits/chosen": -0.06640625,
+      "logits/rejected": -0.1826171875,
+      "logps/chosen": -332.0,
+      "logps/rejected": -300.0,
+      "loss": 0.328,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.75,
+      "step": 441
+    },
+    {
+      "epoch": 0.5423312883435583,
+      "grad_norm": 37.80188895923485,
+      "learning_rate": 6.621487603305785e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3304,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.8125,
+      "step": 442
+    },
+    {
+      "epoch": 0.5435582822085889,
+      "grad_norm": 40.24381216440989,
+      "learning_rate": 6.618181818181818e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3159,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.4140625,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.59375,
+      "step": 443
+    },
+    {
+      "epoch": 0.5447852760736196,
+      "grad_norm": 37.922945446281,
+      "learning_rate": 6.614876033057851e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -446.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2446,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.09375,
+      "rewards/margins": 7.25,
+      "rewards/rejected": -5.15625,
+      "step": 444
+    },
+    {
+      "epoch": 0.5460122699386503,
+      "grad_norm": 38.45325276499507,
+      "learning_rate": 6.611570247933884e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -300.0,
+      "loss": 0.389,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.125,
+      "step": 445
+    },
+    {
+      "epoch": 0.5472392638036809,
+      "grad_norm": 33.02180393783636,
+      "learning_rate": 6.608264462809917e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2348,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.46875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.25,
+      "step": 446
+    },
+    {
+      "epoch": 0.5484662576687117,
+      "grad_norm": 31.670625439038492,
+      "learning_rate": 6.604958677685951e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2551,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 7.15625,
+      "rewards/rejected": -5.28125,
+      "step": 447
+    },
+    {
+      "epoch": 0.5496932515337424,
+      "grad_norm": 37.605046960174505,
+      "learning_rate": 6.601652892561983e-07,
+      "logits/chosen": -0.10888671875,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3209,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.90234375,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -4.25,
+      "step": 448
+    },
+    {
+      "epoch": 0.550920245398773,
+      "grad_norm": 34.080535896094865,
+      "learning_rate": 6.598347107438016e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2712,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -4.53125,
+      "step": 449
+    },
+    {
+      "epoch": 0.5521472392638037,
+      "grad_norm": 35.069217343013015,
+      "learning_rate": 6.595041322314049e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -288.0,
+      "loss": 0.285,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.625,
+      "step": 450
+    },
+    {
+      "epoch": 0.5533742331288344,
+      "grad_norm": 33.87966228532685,
+      "learning_rate": 6.591735537190082e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -340.0,
+      "logps/rejected": -280.0,
+      "loss": 0.3043,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -4.5,
+      "step": 451
+    },
+    {
+      "epoch": 0.554601226993865,
+      "grad_norm": 35.08833886262592,
+      "learning_rate": 6.588429752066115e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -454.0,
+      "logps/rejected": -328.0,
+      "loss": 0.258,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.421875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.21875,
+      "step": 452
+    },
+    {
+      "epoch": 0.5558282208588957,
+      "grad_norm": 30.509116856340853,
+      "learning_rate": 6.585123966942149e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -366.0,
+      "loss": 0.214,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 2.578125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -4.78125,
+      "step": 453
+    },
+    {
+      "epoch": 0.5570552147239264,
+      "grad_norm": 36.06945734305074,
+      "learning_rate": 6.581818181818182e-07,
+      "logits/chosen": -0.11572265625,
+      "logits/rejected": -0.19921875,
+      "logps/chosen": -338.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2945,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.859375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.625,
+      "step": 454
+    },
+    {
+      "epoch": 0.558282208588957,
+      "grad_norm": 35.814869576892626,
+      "learning_rate": 6.578512396694215e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2399,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.9609375,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -4.8125,
+      "step": 455
+    },
+    {
+      "epoch": 0.5595092024539877,
+      "grad_norm": 38.450506141939925,
+      "learning_rate": 6.575206611570248e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.2236328125,
+      "logps/chosen": -312.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2952,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.125,
+      "step": 456
+    },
+    {
+      "epoch": 0.5607361963190184,
+      "grad_norm": 37.93391723235656,
+      "learning_rate": 6.57190082644628e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3025,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.5625,
+      "step": 457
+    },
+    {
+      "epoch": 0.561963190184049,
+      "grad_norm": 33.44669035850576,
+      "learning_rate": 6.568595041322313e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2297,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.125,
+      "step": 458
+    },
+    {
+      "epoch": 0.5631901840490797,
+      "grad_norm": 34.06851635286411,
+      "learning_rate": 6.565289256198347e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2786,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.0625,
+      "step": 459
+    },
+    {
+      "epoch": 0.5644171779141104,
+      "grad_norm": 34.68751260016757,
+      "learning_rate": 6.56198347107438e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -330.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.76171875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -6.03125,
+      "step": 460
+    },
+    {
+      "epoch": 0.5656441717791411,
+      "grad_norm": 38.71078681224242,
+      "learning_rate": 6.558677685950413e-07,
+      "logits/chosen": -0.3515625,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2741,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.34375,
+      "step": 461
+    },
+    {
+      "epoch": 0.5668711656441717,
+      "grad_norm": 39.327400074833584,
+      "learning_rate": 6.555371900826446e-07,
+      "logits/chosen": -0.11279296875,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3248,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.875,
+      "step": 462
+    },
+    {
+      "epoch": 0.5680981595092025,
+      "grad_norm": 37.637164385131356,
+      "learning_rate": 6.552066115702479e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -360.0,
+      "loss": 0.2848,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.98046875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.46875,
+      "step": 463
+    },
+    {
+      "epoch": 0.5693251533742332,
+      "grad_norm": 41.25059009443085,
+      "learning_rate": 6.548760330578512e-07,
+      "logits/chosen": -0.09619140625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2951,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.34375,
+      "step": 464
+    },
+    {
+      "epoch": 0.5705521472392638,
+      "grad_norm": 27.065991136479923,
+      "learning_rate": 6.545454545454546e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -346.0,
+      "loss": 0.1759,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -6.0625,
+      "step": 465
+    },
+    {
+      "epoch": 0.5717791411042945,
+      "grad_norm": 43.40449571309543,
+      "learning_rate": 6.542148760330578e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2899,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.9375,
+      "step": 466
+    },
+    {
+      "epoch": 0.5730061349693252,
+      "grad_norm": 40.392254031751044,
+      "learning_rate": 6.538842975206611e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3087,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.28125,
+      "step": 467
+    },
+    {
+      "epoch": 0.5742331288343558,
+      "grad_norm": 41.32661857185299,
+      "learning_rate": 6.535537190082644e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3022,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.78125,
+      "step": 468
+    },
+    {
+      "epoch": 0.5754601226993865,
+      "grad_norm": 29.07874137818143,
+      "learning_rate": 6.532231404958677e-07,
+      "logits/chosen": -0.12451171875,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1838,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.6875,
+      "step": 469
+    },
+    {
+      "epoch": 0.5766871165644172,
+      "grad_norm": 35.40917214666752,
+      "learning_rate": 6.52892561983471e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.2353515625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -324.0,
+      "loss": 0.301,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.34375,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -5.40625,
+      "step": 470
+    },
+    {
+      "epoch": 0.5779141104294478,
+      "grad_norm": 38.07545863951338,
+      "learning_rate": 6.525619834710744e-07,
+      "logits/chosen": -0.03369140625,
+      "logits/rejected": -0.203125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3263,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -5.125,
+      "step": 471
+    },
+    {
+      "epoch": 0.5791411042944785,
+      "grad_norm": 40.21918996431336,
+      "learning_rate": 6.522314049586777e-07,
+      "logits/chosen": -0.08447265625,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3173,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -5.15625,
+      "step": 472
+    },
+    {
+      "epoch": 0.5803680981595092,
+      "grad_norm": 37.03158250739783,
+      "learning_rate": 6.51900826446281e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -378.0,
+      "loss": 0.3328,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.90625,
+      "step": 473
+    },
+    {
+      "epoch": 0.5815950920245399,
+      "grad_norm": 38.12266289452875,
+      "learning_rate": 6.515702479338843e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -430.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2648,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -5.21875,
+      "step": 474
+    },
+    {
+      "epoch": 0.5828220858895705,
+      "grad_norm": 31.08159067757818,
+      "learning_rate": 6.512396694214875e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2361,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 7.59375,
+      "rewards/rejected": -5.59375,
+      "step": 475
+    },
+    {
+      "epoch": 0.5840490797546012,
+      "grad_norm": 31.0829148956811,
+      "learning_rate": 6.509090909090908e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2813,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.185546875,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -5.6875,
+      "step": 476
+    },
+    {
+      "epoch": 0.5852760736196319,
+      "grad_norm": 36.7408304979899,
+      "learning_rate": 6.505785123966942e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -354.0,
+      "loss": 0.257,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.75,
+      "step": 477
+    },
+    {
+      "epoch": 0.5865030674846625,
+      "grad_norm": 34.872893059723594,
+      "learning_rate": 6.502479338842975e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.84375,
+      "step": 478
+    },
+    {
+      "epoch": 0.5877300613496933,
+      "grad_norm": 35.26754811760802,
+      "learning_rate": 6.499173553719008e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2784,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.240234375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.96875,
+      "step": 479
+    },
+    {
+      "epoch": 0.588957055214724,
+      "grad_norm": 33.02148763705059,
+      "learning_rate": 6.495867768595041e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2364,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.4765625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -6.125,
+      "step": 480
+    },
+    {
+      "epoch": 0.5901840490797546,
+      "grad_norm": 39.97994667673503,
+      "learning_rate": 6.492561983471074e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -378.0,
+      "loss": 0.3732,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.84765625,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.46875,
+      "step": 481
+    },
+    {
+      "epoch": 0.5914110429447853,
+      "grad_norm": 33.72401411710678,
+      "learning_rate": 6.489256198347108e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2739,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.4375,
+      "step": 482
+    },
+    {
+      "epoch": 0.592638036809816,
+      "grad_norm": 34.19291475160813,
+      "learning_rate": 6.485950413223141e-07,
+      "logits/chosen": -0.0703125,
+      "logits/rejected": -0.2138671875,
+      "logps/chosen": -346.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2917,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.5625,
+      "step": 483
+    },
+    {
+      "epoch": 0.5938650306748466,
+      "grad_norm": 35.398942571150606,
+      "learning_rate": 6.482644628099173e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3155,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.1875,
+      "rewards/rejected": -4.875,
+      "step": 484
+    },
+    {
+      "epoch": 0.5950920245398773,
+      "grad_norm": 30.26872146503766,
+      "learning_rate": 6.479338842975206e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -306.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.625,
+      "step": 485
+    },
+    {
+      "epoch": 0.596319018404908,
+      "grad_norm": 35.38748335338782,
+      "learning_rate": 6.476033057851239e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2658,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -4.96875,
+      "step": 486
+    },
+    {
+      "epoch": 0.5975460122699386,
+      "grad_norm": 30.285435869887902,
+      "learning_rate": 6.472727272727272e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -350.0,
+      "loss": 0.233,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.5625,
+      "step": 487
+    },
+    {
+      "epoch": 0.5987730061349693,
+      "grad_norm": 37.89908939404138,
+      "learning_rate": 6.469421487603306e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2436,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.0625,
+      "step": 488
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 30.515429632482274,
+      "learning_rate": 6.466115702479339e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -326.0,
+      "loss": 0.272,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 7.3125,
+      "rewards/rejected": -5.5,
+      "step": 489
+    },
+    {
+      "epoch": 0.6012269938650306,
+      "grad_norm": 40.37971138033072,
+      "learning_rate": 6.462809917355372e-07,
+      "logits/chosen": -0.09765625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3004,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -5.1875,
+      "step": 490
+    },
+    {
+      "epoch": 0.6024539877300613,
+      "grad_norm": 34.45143017996783,
+      "learning_rate": 6.459504132231405e-07,
+      "logits/chosen": -0.08203125,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2549,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.4375,
+      "step": 491
+    },
+    {
+      "epoch": 0.603680981595092,
+      "grad_norm": 36.58265725949472,
+      "learning_rate": 6.456198347107438e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -378.0,
+      "loss": 0.2737,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.9375,
+      "step": 492
+    },
+    {
+      "epoch": 0.6049079754601226,
+      "grad_norm": 42.46481332966053,
+      "learning_rate": 6.45289256198347e-07,
+      "logits/chosen": -0.024169921875,
+      "logits/rejected": -0.134765625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3917,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -4.25,
+      "step": 493
+    },
+    {
+      "epoch": 0.6061349693251534,
+      "grad_norm": 33.59404853921049,
+      "learning_rate": 6.449586776859504e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2643,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.09375,
+      "step": 494
+    },
+    {
+      "epoch": 0.6073619631901841,
+      "grad_norm": 37.484469870256234,
+      "learning_rate": 6.446280991735537e-07,
+      "logits/chosen": -0.1220703125,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3066,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.21875,
+      "step": 495
+    },
+    {
+      "epoch": 0.6085889570552148,
+      "grad_norm": 40.63137683640751,
+      "learning_rate": 6.44297520661157e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3177,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -5.28125,
+      "step": 496
+    },
+    {
+      "epoch": 0.6098159509202454,
+      "grad_norm": 41.064298986990785,
+      "learning_rate": 6.439669421487603e-07,
+      "logits/chosen": -0.12158203125,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3865,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.85546875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.5,
+      "step": 497
+    },
+    {
+      "epoch": 0.6110429447852761,
+      "grad_norm": 34.6148352569408,
+      "learning_rate": 6.436363636363636e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -358.0,
+      "loss": 0.2817,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.65625,
+      "step": 498
+    },
+    {
+      "epoch": 0.6122699386503068,
+      "grad_norm": 36.60546803390198,
+      "learning_rate": 6.433057851239669e-07,
+      "logits/chosen": -0.115234375,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2842,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.125,
+      "step": 499
+    },
+    {
+      "epoch": 0.6134969325153374,
+      "grad_norm": 33.21675862161022,
+      "learning_rate": 6.429752066115703e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2718,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.21875,
+      "step": 500
+    },
+    {
+      "epoch": 0.6147239263803681,
+      "grad_norm": 32.544310987499145,
+      "learning_rate": 6.426446280991735e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -344.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.625,
+      "step": 501
+    },
+    {
+      "epoch": 0.6159509202453988,
+      "grad_norm": 32.09170254227315,
+      "learning_rate": 6.423140495867768e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2639,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 7.46875,
+      "rewards/rejected": -5.96875,
+      "step": 502
+    },
+    {
+      "epoch": 0.6171779141104294,
+      "grad_norm": 35.53103605940553,
+      "learning_rate": 6.419834710743801e-07,
+      "logits/chosen": 0.01153564453125,
+      "logits/rejected": -0.1904296875,
+      "logps/chosen": -326.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2795,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.25,
+      "step": 503
+    },
+    {
+      "epoch": 0.6184049079754601,
+      "grad_norm": 30.299429390734634,
+      "learning_rate": 6.416528925619834e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2658,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -5.40625,
+      "step": 504
+    },
+    {
+      "epoch": 0.6196319018404908,
+      "grad_norm": 44.689012762744575,
+      "learning_rate": 6.413223140495867e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -394.0,
+      "loss": 0.3776,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.369140625,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -6.15625,
+      "step": 505
+    },
+    {
+      "epoch": 0.6208588957055214,
+      "grad_norm": 32.33588195685234,
+      "learning_rate": 6.409917355371901e-07,
+      "logits/chosen": -0.08349609375,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2588,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.90625,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.875,
+      "step": 506
+    },
+    {
+      "epoch": 0.6220858895705521,
+      "grad_norm": 29.030680879470466,
+      "learning_rate": 6.406611570247934e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -308.0,
+      "loss": 0.1953,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.90234375,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -6.15625,
+      "step": 507
+    },
+    {
+      "epoch": 0.6233128834355828,
+      "grad_norm": 31.4626526695897,
+      "learning_rate": 6.403305785123967e-07,
+      "logits/chosen": -0.0201416015625,
+      "logits/rejected": -0.2021484375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -310.0,
+      "loss": 0.24,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.671875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.59375,
+      "step": 508
+    },
+    {
+      "epoch": 0.6245398773006134,
+      "grad_norm": 42.61520515559976,
+      "learning_rate": 6.4e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -466.0,
+      "logps/rejected": -384.0,
+      "loss": 0.3973,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8984375,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -6.0625,
+      "step": 509
+    },
+    {
+      "epoch": 0.6257668711656442,
+      "grad_norm": 30.165060750168916,
+      "learning_rate": 6.396694214876032e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2226,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.306640625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.625,
+      "step": 510
+    },
+    {
+      "epoch": 0.6269938650306749,
+      "grad_norm": 40.15670022385489,
+      "learning_rate": 6.393388429752065e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3348,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.81640625,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -6.1875,
+      "step": 511
+    },
+    {
+      "epoch": 0.6282208588957056,
+      "grad_norm": 44.290408519253646,
+      "learning_rate": 6.390082644628099e-07,
+      "logits/chosen": -0.09130859375,
+      "logits/rejected": -0.2431640625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -384.0,
+      "loss": 0.3621,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.35546875,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -5.0,
+      "step": 512
+    },
+    {
+      "epoch": 0.6294478527607362,
+      "grad_norm": 37.44250565144745,
+      "learning_rate": 6.386776859504132e-07,
+      "logits/chosen": -0.09521484375,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -312.0,
+      "loss": 0.337,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -5.375,
+      "step": 513
+    },
+    {
+      "epoch": 0.6306748466257669,
+      "grad_norm": 29.97245642529846,
+      "learning_rate": 6.383471074380165e-07,
+      "logits/chosen": -0.0830078125,
+      "logits/rejected": -0.2412109375,
+      "logps/chosen": -330.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2436,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": -0.051513671875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -6.03125,
+      "step": 514
+    },
+    {
+      "epoch": 0.6319018404907976,
+      "grad_norm": 36.989875573277935,
+      "learning_rate": 6.380165289256198e-07,
+      "logits/chosen": -0.09912109375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2882,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": -0.041748046875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.84375,
+      "step": 515
+    },
+    {
+      "epoch": 0.6331288343558282,
+      "grad_norm": 31.306859511596812,
+      "learning_rate": 6.376859504132231e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2432,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.1142578125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.84375,
+      "step": 516
+    },
+    {
+      "epoch": 0.6343558282208589,
+      "grad_norm": 43.67170294973452,
+      "learning_rate": 6.373553719008264e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3884,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.53125,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -5.03125,
+      "step": 517
+    },
+    {
+      "epoch": 0.6355828220858896,
+      "grad_norm": 29.354877414703584,
+      "learning_rate": 6.370247933884298e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -342.0,
+      "loss": 0.1963,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.34375,
+      "step": 518
+    },
+    {
+      "epoch": 0.6368098159509202,
+      "grad_norm": 40.61854272265548,
+      "learning_rate": 6.36694214876033e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -334.0,
+      "loss": 0.299,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.75390625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.5,
+      "step": 519
+    },
+    {
+      "epoch": 0.6380368098159509,
+      "grad_norm": 38.052325639181014,
+      "learning_rate": 6.363636363636363e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3436,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.6875,
+      "step": 520
+    },
+    {
+      "epoch": 0.6392638036809816,
+      "grad_norm": 32.37002227141111,
+      "learning_rate": 6.360330578512396e-07,
+      "logits/chosen": -0.03271484375,
+      "logits/rejected": -0.171875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2525,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.5,
+      "step": 521
+    },
+    {
+      "epoch": 0.6404907975460122,
+      "grad_norm": 40.41274539784558,
+      "learning_rate": 6.357024793388429e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -420.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3152,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.96875,
+      "step": 522
+    },
+    {
+      "epoch": 0.6417177914110429,
+      "grad_norm": 42.30362208986331,
+      "learning_rate": 6.353719008264462e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3605,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.6171875,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -5.03125,
+      "step": 523
+    },
+    {
+      "epoch": 0.6429447852760736,
+      "grad_norm": 46.893532525519284,
+      "learning_rate": 6.350413223140496e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3361,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.0625,
+      "step": 524
+    },
+    {
+      "epoch": 0.6441717791411042,
+      "grad_norm": 49.422180518954264,
+      "learning_rate": 6.347107438016529e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.185546875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3927,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.498046875,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -4.65625,
+      "step": 525
+    },
+    {
+      "epoch": 0.645398773006135,
+      "grad_norm": 38.5530601675623,
+      "learning_rate": 6.343801652892562e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2504,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.5625,
+      "step": 526
+    },
+    {
+      "epoch": 0.6466257668711657,
+      "grad_norm": 35.94650372227145,
+      "learning_rate": 6.340495867768595e-07,
+      "logits/chosen": -0.0859375,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2888,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.9609375,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.1875,
+      "step": 527
+    },
+    {
+      "epoch": 0.6478527607361964,
+      "grad_norm": 33.57585022481078,
+      "learning_rate": 6.337190082644627e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2986,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.71875,
+      "step": 528
+    },
+    {
+      "epoch": 0.649079754601227,
+      "grad_norm": 35.863933883867006,
+      "learning_rate": 6.33388429752066e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3657,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.1875,
+      "step": 529
+    },
+    {
+      "epoch": 0.6503067484662577,
+      "grad_norm": 41.86945425639466,
+      "learning_rate": 6.330578512396694e-07,
+      "logits/chosen": -0.12353515625,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3197,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.375,
+      "step": 530
+    },
+    {
+      "epoch": 0.6515337423312884,
+      "grad_norm": 33.3934976525138,
+      "learning_rate": 6.327272727272727e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2751,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -5.1875,
+      "step": 531
+    },
+    {
+      "epoch": 0.652760736196319,
+      "grad_norm": 35.72821914002383,
+      "learning_rate": 6.32396694214876e-07,
+      "logits/chosen": -0.07958984375,
+      "logits/rejected": -0.2197265625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2782,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.875,
+      "step": 532
+    },
+    {
+      "epoch": 0.6539877300613497,
+      "grad_norm": 36.05732236551947,
+      "learning_rate": 6.320661157024793e-07,
+      "logits/chosen": -0.2421875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2611,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.8671875,
+      "rewards/margins": 7.3125,
+      "rewards/rejected": -5.4375,
+      "step": 533
+    },
+    {
+      "epoch": 0.6552147239263804,
+      "grad_norm": 43.004226294889776,
+      "learning_rate": 6.317355371900826e-07,
+      "logits/chosen": -0.10302734375,
+      "logits/rejected": -0.228515625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2813,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.84375,
+      "step": 534
+    },
+    {
+      "epoch": 0.656441717791411,
+      "grad_norm": 39.5682500491772,
+      "learning_rate": 6.31404958677686e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2961,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.380859375,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.5625,
+      "step": 535
+    },
+    {
+      "epoch": 0.6576687116564417,
+      "grad_norm": 36.55745143954986,
+      "learning_rate": 6.310743801652893e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2275,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.4375,
+      "step": 536
+    },
+    {
+      "epoch": 0.6588957055214724,
+      "grad_norm": 45.10266990819029,
+      "learning_rate": 6.307438016528925e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3939,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.265625,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -4.84375,
+      "step": 537
+    },
+    {
+      "epoch": 0.660122699386503,
+      "grad_norm": 44.09458580737532,
+      "learning_rate": 6.304132231404958e-07,
+      "logits/chosen": -0.1123046875,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3783,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.84375,
+      "step": 538
+    },
+    {
+      "epoch": 0.6613496932515337,
+      "grad_norm": 36.16672414551654,
+      "learning_rate": 6.300826446280991e-07,
+      "logits/chosen": -0.11181640625,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -340.0,
+      "loss": 0.309,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.40625,
+      "step": 539
+    },
+    {
+      "epoch": 0.6625766871165644,
+      "grad_norm": 39.077710779231424,
+      "learning_rate": 6.297520661157024e-07,
+      "logits/chosen": -0.0908203125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3217,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 6.65625,
+      "rewards/rejected": -5.15625,
+      "step": 540
+    },
+    {
+      "epoch": 0.6638036809815951,
+      "grad_norm": 35.971169171679556,
+      "learning_rate": 6.294214876033058e-07,
+      "logits/chosen": -0.10986328125,
+      "logits/rejected": -0.2080078125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2834,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.75390625,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -4.8125,
+      "step": 541
+    },
+    {
+      "epoch": 0.6650306748466258,
+      "grad_norm": 34.168657597082216,
+      "learning_rate": 6.290909090909091e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2812,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.65625,
+      "step": 542
+    },
+    {
+      "epoch": 0.6662576687116565,
+      "grad_norm": 36.23729440291612,
+      "learning_rate": 6.287603305785124e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -444.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2622,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -4.78125,
+      "step": 543
+    },
+    {
+      "epoch": 0.6674846625766871,
+      "grad_norm": 45.238447000820585,
+      "learning_rate": 6.284297520661157e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3695,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.91015625,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.25,
+      "step": 544
+    },
+    {
+      "epoch": 0.6687116564417178,
+      "grad_norm": 43.061817496062964,
+      "learning_rate": 6.28099173553719e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -452.0,
+      "logps/rejected": -360.0,
+      "loss": 0.4047,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.5625,
+      "step": 545
+    },
+    {
+      "epoch": 0.6699386503067485,
+      "grad_norm": 40.8862609592416,
+      "learning_rate": 6.277685950413222e-07,
+      "logits/chosen": -0.10205078125,
+      "logits/rejected": -0.20703125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3402,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.09375,
+      "step": 546
+    },
+    {
+      "epoch": 0.6711656441717792,
+      "grad_norm": 32.99712535949366,
+      "learning_rate": 6.274380165289256e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2883,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.9375,
+      "step": 547
+    },
+    {
+      "epoch": 0.6723926380368098,
+      "grad_norm": 40.33054062017028,
+      "learning_rate": 6.271074380165289e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2583,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.71875,
+      "step": 548
+    },
+    {
+      "epoch": 0.6736196319018405,
+      "grad_norm": 37.72507019090557,
+      "learning_rate": 6.267768595041322e-07,
+      "logits/chosen": -0.11474609375,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.84375,
+      "step": 549
+    },
+    {
+      "epoch": 0.6748466257668712,
+      "grad_norm": 37.30173963897267,
+      "learning_rate": 6.264462809917355e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3323,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.59375,
+      "step": 550
+    },
+    {
+      "epoch": 0.6760736196319018,
+      "grad_norm": 33.242769441390706,
+      "learning_rate": 6.261157024793388e-07,
+      "logits/chosen": -0.0986328125,
+      "logits/rejected": -0.2265625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2814,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9296875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -4.59375,
+      "step": 551
+    },
+    {
+      "epoch": 0.6773006134969325,
+      "grad_norm": 35.53857403218536,
+      "learning_rate": 6.257851239669421e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3055,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.6875,
+      "step": 552
+    },
+    {
+      "epoch": 0.6785276073619632,
+      "grad_norm": 39.8186178694702,
+      "learning_rate": 6.254545454545455e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3556,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.8125,
+      "step": 553
+    },
+    {
+      "epoch": 0.6797546012269938,
+      "grad_norm": 34.396570445757696,
+      "learning_rate": 6.251239669421488e-07,
+      "logits/chosen": -0.08349609375,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -326.0,
+      "loss": 0.261,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.6875,
+      "step": 554
+    },
+    {
+      "epoch": 0.6809815950920245,
+      "grad_norm": 37.46750845748253,
+      "learning_rate": 6.24793388429752e-07,
+      "logits/chosen": 0.0196533203125,
+      "logits/rejected": -0.107421875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3763,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.65625,
+      "step": 555
+    },
+    {
+      "epoch": 0.6822085889570552,
+      "grad_norm": 28.75095679008507,
+      "learning_rate": 6.244628099173553e-07,
+      "logits/chosen": -0.10693359375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2294,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.0625,
+      "step": 556
+    },
+    {
+      "epoch": 0.6834355828220859,
+      "grad_norm": 32.12117772558378,
+      "learning_rate": 6.241322314049586e-07,
+      "logits/chosen": -0.11669921875,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2775,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -4.65625,
+      "step": 557
+    },
+    {
+      "epoch": 0.6846625766871166,
+      "grad_norm": 30.45940478111348,
+      "learning_rate": 6.238016528925619e-07,
+      "logits/chosen": -0.07861328125,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2442,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.8359375,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.875,
+      "step": 558
+    },
+    {
+      "epoch": 0.6858895705521473,
+      "grad_norm": 38.28759745643993,
+      "learning_rate": 6.234710743801653e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3234,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.25,
+      "step": 559
+    },
+    {
+      "epoch": 0.6871165644171779,
+      "grad_norm": 25.47174077723704,
+      "learning_rate": 6.231404958677686e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2016,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -6.03125,
+      "step": 560
+    },
+    {
+      "epoch": 0.6883435582822086,
+      "grad_norm": 28.61531386213562,
+      "learning_rate": 6.228099173553719e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -300.0,
+      "loss": 0.2021,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.6875,
+      "step": 561
+    },
+    {
+      "epoch": 0.6895705521472393,
+      "grad_norm": 30.528229373220828,
+      "learning_rate": 6.224793388429752e-07,
+      "logits/chosen": -0.032470703125,
+      "logits/rejected": -0.2333984375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2462,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.5,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.5625,
+      "step": 562
+    },
+    {
+      "epoch": 0.69079754601227,
+      "grad_norm": 34.031991776353316,
+      "learning_rate": 6.221487603305785e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2904,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 7.21875,
+      "rewards/rejected": -5.65625,
+      "step": 563
+    },
+    {
+      "epoch": 0.6920245398773006,
+      "grad_norm": 24.836619890425215,
+      "learning_rate": 6.218181818181817e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -334.0,
+      "loss": 0.224,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.34375,
+      "step": 564
+    },
+    {
+      "epoch": 0.6932515337423313,
+      "grad_norm": 35.09286389818353,
+      "learning_rate": 6.214876033057851e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3639,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.6015625,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.84375,
+      "step": 565
+    },
+    {
+      "epoch": 0.694478527607362,
+      "grad_norm": 33.71563020260373,
+      "learning_rate": 6.211570247933884e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2462,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.1875,
+      "step": 566
+    },
+    {
+      "epoch": 0.6957055214723926,
+      "grad_norm": 25.84647564741135,
+      "learning_rate": 6.208264462809917e-07,
+      "logits/chosen": -0.053955078125,
+      "logits/rejected": -0.1875,
+      "logps/chosen": -316.0,
+      "logps/rejected": -274.0,
+      "loss": 0.2061,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.359375,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -5.09375,
+      "step": 567
+    },
+    {
+      "epoch": 0.6969325153374233,
+      "grad_norm": 37.567177147049684,
+      "learning_rate": 6.20495867768595e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3064,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.21875,
+      "step": 568
+    },
+    {
+      "epoch": 0.698159509202454,
+      "grad_norm": 32.35117698678835,
+      "learning_rate": 6.201652892561983e-07,
+      "logits/chosen": -0.10302734375,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2508,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.125,
+      "step": 569
+    },
+    {
+      "epoch": 0.6993865030674846,
+      "grad_norm": 25.32334987337483,
+      "learning_rate": 6.198347107438017e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -338.0,
+      "loss": 0.1649,
+      "rewards/accuracies": 0.9453125,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.34375,
+      "step": 570
+    },
+    {
+      "epoch": 0.7006134969325153,
+      "grad_norm": 31.066523168854104,
+      "learning_rate": 6.19504132231405e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -326.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2262,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.6015625,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -5.09375,
+      "step": 571
+    },
+    {
+      "epoch": 0.701840490797546,
+      "grad_norm": 32.881821449548816,
+      "learning_rate": 6.191735537190083e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2291,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -4.9375,
+      "step": 572
+    },
+    {
+      "epoch": 0.7030674846625767,
+      "grad_norm": 40.68529761937124,
+      "learning_rate": 6.188429752066115e-07,
+      "logits/chosen": -0.07177734375,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4034,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -5.3125,
+      "step": 573
+    },
+    {
+      "epoch": 0.7042944785276074,
+      "grad_norm": 38.820291986225,
+      "learning_rate": 6.185123966942148e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -322.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2792,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.25,
+      "step": 574
+    },
+    {
+      "epoch": 0.7055214723926381,
+      "grad_norm": 33.59717634578736,
+      "learning_rate": 6.181818181818181e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2284,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.59375,
+      "step": 575
+    },
+    {
+      "epoch": 0.7067484662576687,
+      "grad_norm": 28.04916589516562,
+      "learning_rate": 6.178512396694215e-07,
+      "logits/chosen": -0.10302734375,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2001,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -5.875,
+      "step": 576
+    },
+    {
+      "epoch": 0.7079754601226994,
+      "grad_norm": 36.15869255614372,
+      "learning_rate": 6.175206611570248e-07,
+      "logits/chosen": -0.03955078125,
+      "logits/rejected": -0.1767578125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2929,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.84375,
+      "step": 577
+    },
+    {
+      "epoch": 0.7092024539877301,
+      "grad_norm": 36.89994352340565,
+      "learning_rate": 6.171900826446281e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -334.0,
+      "loss": 0.322,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.75,
+      "step": 578
+    },
+    {
+      "epoch": 0.7104294478527607,
+      "grad_norm": 42.55492037495249,
+      "learning_rate": 6.168595041322314e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3864,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.96875,
+      "step": 579
+    },
+    {
+      "epoch": 0.7116564417177914,
+      "grad_norm": 35.02868013813214,
+      "learning_rate": 6.165289256198347e-07,
+      "logits/chosen": -0.07275390625,
+      "logits/rejected": -0.2451171875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -354.0,
+      "loss": 0.229,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.73046875,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -6.25,
+      "step": 580
+    },
+    {
+      "epoch": 0.7128834355828221,
+      "grad_norm": 29.748844282456343,
+      "learning_rate": 6.16198347107438e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2071,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.84375,
+      "step": 581
+    },
+    {
+      "epoch": 0.7141104294478527,
+      "grad_norm": 41.261123421442086,
+      "learning_rate": 6.158677685950413e-07,
+      "logits/chosen": 0.01953125,
+      "logits/rejected": -0.197265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3132,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.78125,
+      "step": 582
+    },
+    {
+      "epoch": 0.7153374233128834,
+      "grad_norm": 37.14443132697755,
+      "learning_rate": 6.155371900826446e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3036,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.875,
+      "step": 583
+    },
+    {
+      "epoch": 0.7165644171779141,
+      "grad_norm": 42.810461239623855,
+      "learning_rate": 6.152066115702479e-07,
+      "logits/chosen": -0.0693359375,
+      "logits/rejected": -0.14453125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -370.0,
+      "loss": 0.3929,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 0.5859375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.25,
+      "step": 584
+    },
+    {
+      "epoch": 0.7177914110429447,
+      "grad_norm": 44.860424166016166,
+      "learning_rate": 6.148760330578512e-07,
+      "logits/chosen": 0.0206298828125,
+      "logits/rejected": -0.099609375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -322.0,
+      "loss": 0.4253,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.0625,
+      "step": 585
+    },
+    {
+      "epoch": 0.7190184049079754,
+      "grad_norm": 36.65848945690881,
+      "learning_rate": 6.145454545454545e-07,
+      "logits/chosen": -0.1064453125,
+      "logits/rejected": -0.2158203125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3226,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.5625,
+      "step": 586
+    },
+    {
+      "epoch": 0.7202453987730061,
+      "grad_norm": 34.66555869797275,
+      "learning_rate": 6.142148760330578e-07,
+      "logits/chosen": -0.1396484375,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -332.0,
+      "logps/rejected": -308.0,
+      "loss": 0.319,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.5,
+      "step": 587
+    },
+    {
+      "epoch": 0.7214723926380369,
+      "grad_norm": 25.81313251350675,
+      "learning_rate": 6.138842975206612e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2328,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9765625,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -4.34375,
+      "step": 588
+    },
+    {
+      "epoch": 0.7226993865030675,
+      "grad_norm": 39.45475953492192,
+      "learning_rate": 6.135537190082645e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3428,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.7734375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.0625,
+      "step": 589
+    },
+    {
+      "epoch": 0.7239263803680982,
+      "grad_norm": 28.739570926494295,
+      "learning_rate": 6.132231404958677e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.2392578125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2629,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -4.375,
+      "step": 590
+    },
+    {
+      "epoch": 0.7251533742331289,
+      "grad_norm": 39.542558184509424,
+      "learning_rate": 6.12892561983471e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2455,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -4.9375,
+      "step": 591
+    },
+    {
+      "epoch": 0.7263803680981595,
+      "grad_norm": 31.700565699717185,
+      "learning_rate": 6.125619834710743e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2372,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.1875,
+      "step": 592
+    },
+    {
+      "epoch": 0.7276073619631902,
+      "grad_norm": 43.160388567583965,
+      "learning_rate": 6.122314049586776e-07,
+      "logits/chosen": -0.08251953125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -442.0,
+      "logps/rejected": -366.0,
+      "loss": 0.3556,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -4.5625,
+      "step": 593
+    },
+    {
+      "epoch": 0.7288343558282209,
+      "grad_norm": 33.17336982955441,
+      "learning_rate": 6.11900826446281e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2737,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.59375,
+      "step": 594
+    },
+    {
+      "epoch": 0.7300613496932515,
+      "grad_norm": 35.78921531146286,
+      "learning_rate": 6.115702479338843e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -444.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2188,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.40625,
+      "step": 595
+    },
+    {
+      "epoch": 0.7312883435582822,
+      "grad_norm": 39.132566907776244,
+      "learning_rate": 6.112396694214876e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3258,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.125,
+      "step": 596
+    },
+    {
+      "epoch": 0.7325153374233129,
+      "grad_norm": 36.16852227530472,
+      "learning_rate": 6.109090909090909e-07,
+      "logits/chosen": -0.03515625,
+      "logits/rejected": -0.2138671875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3225,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.9375,
+      "step": 597
+    },
+    {
+      "epoch": 0.7337423312883435,
+      "grad_norm": 29.88526191086389,
+      "learning_rate": 6.105785123966942e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -438.0,
+      "logps/rejected": -318.0,
+      "loss": 0.204,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 7.21875,
+      "rewards/rejected": -5.125,
+      "step": 598
+    },
+    {
+      "epoch": 0.7349693251533742,
+      "grad_norm": 32.885582252909025,
+      "learning_rate": 6.102479338842974e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -448.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2847,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.53125,
+      "step": 599
+    },
+    {
+      "epoch": 0.7361963190184049,
+      "grad_norm": 37.722786017646094,
+      "learning_rate": 6.099173553719008e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -318.0,
+      "loss": 0.368,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.53125,
+      "step": 600
+    },
+    {
+      "epoch": 0.7374233128834355,
+      "grad_norm": 31.64488245915107,
+      "learning_rate": 6.095867768595041e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2698,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.1826171875,
+      "rewards/margins": 6.125,
+      "rewards/rejected": -5.9375,
+      "step": 601
+    },
+    {
+      "epoch": 0.7386503067484662,
+      "grad_norm": 25.124092304435703,
+      "learning_rate": 6.092561983471074e-07,
+      "logits/chosen": -0.0076904296875,
+      "logits/rejected": -0.1708984375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -332.0,
+      "loss": 0.189,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 7.28125,
+      "rewards/rejected": -6.03125,
+      "step": 602
+    },
+    {
+      "epoch": 0.7398773006134969,
+      "grad_norm": 30.45546682074925,
+      "learning_rate": 6.089256198347107e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2864,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.625,
+      "step": 603
+    },
+    {
+      "epoch": 0.7411042944785277,
+      "grad_norm": 31.94436328392251,
+      "learning_rate": 6.08595041322314e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2464,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.90625,
+      "step": 604
+    },
+    {
+      "epoch": 0.7423312883435583,
+      "grad_norm": 33.65745564504924,
+      "learning_rate": 6.082644628099173e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2598,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.984375,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -6.15625,
+      "step": 605
+    },
+    {
+      "epoch": 0.743558282208589,
+      "grad_norm": 38.63161562009038,
+      "learning_rate": 6.079338842975207e-07,
+      "logits/chosen": -0.2421875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3454,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.3125,
+      "step": 606
+    },
+    {
+      "epoch": 0.7447852760736197,
+      "grad_norm": 33.230355625984764,
+      "learning_rate": 6.07603305785124e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2611,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.48828125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -6.40625,
+      "step": 607
+    },
+    {
+      "epoch": 0.7460122699386503,
+      "grad_norm": 33.50647832359134,
+      "learning_rate": 6.072727272727272e-07,
+      "logits/chosen": -0.0216064453125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2913,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.099609375,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -5.6875,
+      "step": 608
+    },
+    {
+      "epoch": 0.747239263803681,
+      "grad_norm": 35.459553964773974,
+      "learning_rate": 6.069421487603305e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2856,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.54296875,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -6.21875,
+      "step": 609
+    },
+    {
+      "epoch": 0.7484662576687117,
+      "grad_norm": 31.662560042647854,
+      "learning_rate": 6.066115702479338e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -444.0,
+      "logps/rejected": -376.0,
+      "loss": 0.2452,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 7.46875,
+      "rewards/rejected": -6.375,
+      "step": 610
+    },
+    {
+      "epoch": 0.7496932515337423,
+      "grad_norm": 49.54193875289406,
+      "learning_rate": 6.062809917355371e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -362.0,
+      "loss": 0.3976,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -0.1396484375,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -6.0,
+      "step": 611
+    },
+    {
+      "epoch": 0.750920245398773,
+      "grad_norm": 39.062720193674714,
+      "learning_rate": 6.059504132231405e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3777,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.546875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.40625,
+      "step": 612
+    },
+    {
+      "epoch": 0.7521472392638037,
+      "grad_norm": 35.9253243800041,
+      "learning_rate": 6.056198347107438e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.2216796875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3394,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.10107421875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.96875,
+      "step": 613
+    },
+    {
+      "epoch": 0.7533742331288343,
+      "grad_norm": 54.139929990418025,
+      "learning_rate": 6.052892561983471e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3142,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.78125,
+      "step": 614
+    },
+    {
+      "epoch": 0.754601226993865,
+      "grad_norm": 37.247633297004576,
+      "learning_rate": 6.049586776859504e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3089,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.84765625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -6.0,
+      "step": 615
+    },
+    {
+      "epoch": 0.7558282208588957,
+      "grad_norm": 31.187596328354086,
+      "learning_rate": 6.046280991735537e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2528,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.10302734375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -5.46875,
+      "step": 616
+    },
+    {
+      "epoch": 0.7570552147239263,
+      "grad_norm": 39.842219715736135,
+      "learning_rate": 6.04297520661157e-07,
+      "logits/chosen": -0.125,
+      "logits/rejected": -0.208984375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3542,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.48828125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.3125,
+      "step": 617
+    },
+    {
+      "epoch": 0.758282208588957,
+      "grad_norm": 43.96609425022703,
+      "learning_rate": 6.039669421487603e-07,
+      "logits/chosen": -0.052734375,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3723,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.51953125,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.9375,
+      "step": 618
+    },
+    {
+      "epoch": 0.7595092024539877,
+      "grad_norm": 32.587670722826026,
+      "learning_rate": 6.036363636363636e-07,
+      "logits/chosen": -0.0791015625,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2331,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.90625,
+      "step": 619
+    },
+    {
+      "epoch": 0.7607361963190185,
+      "grad_norm": 44.41642388280846,
+      "learning_rate": 6.033057851239669e-07,
+      "logits/chosen": -0.0830078125,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3088,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.5234375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.21875,
+      "step": 620
+    },
+    {
+      "epoch": 0.7619631901840491,
+      "grad_norm": 34.02108964551065,
+      "learning_rate": 6.029752066115702e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -478.0,
+      "logps/rejected": -368.0,
+      "loss": 0.2661,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 7.3125,
+      "rewards/rejected": -6.125,
+      "step": 621
+    },
+    {
+      "epoch": 0.7631901840490798,
+      "grad_norm": 31.832438088278238,
+      "learning_rate": 6.026446280991735e-07,
+      "logits/chosen": -0.0888671875,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -342.0,
+      "loss": 0.236,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.72265625,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -6.03125,
+      "step": 622
+    },
+    {
+      "epoch": 0.7644171779141105,
+      "grad_norm": 38.89511615880052,
+      "learning_rate": 6.023140495867769e-07,
+      "logits/chosen": -0.1181640625,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -452.0,
+      "logps/rejected": -376.0,
+      "loss": 0.3167,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.4453125,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.46875,
+      "step": 623
+    },
+    {
+      "epoch": 0.7656441717791411,
+      "grad_norm": 32.19915482428781,
+      "learning_rate": 6.019834710743802e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2307,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 7.78125,
+      "rewards/rejected": -6.53125,
+      "step": 624
+    },
+    {
+      "epoch": 0.7668711656441718,
+      "grad_norm": 32.198949298871604,
+      "learning_rate": 6.016528925619835e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2191,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.375,
+      "step": 625
+    },
+    {
+      "epoch": 0.7680981595092025,
+      "grad_norm": 40.391116665536885,
+      "learning_rate": 6.013223140495867e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3346,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.03125,
+      "step": 626
+    },
+    {
+      "epoch": 0.7693251533742331,
+      "grad_norm": 33.50213686948866,
+      "learning_rate": 6.0099173553719e-07,
+      "logits/chosen": -0.06396484375,
+      "logits/rejected": -0.154296875,
+      "logps/chosen": -310.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3297,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": -0.1552734375,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -5.21875,
+      "step": 627
+    },
+    {
+      "epoch": 0.7705521472392638,
+      "grad_norm": 37.25258333430096,
+      "learning_rate": 6.006611570247933e-07,
+      "logits/chosen": -0.0947265625,
+      "logits/rejected": -0.2392578125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2886,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.90625,
+      "step": 628
+    },
+    {
+      "epoch": 0.7717791411042945,
+      "grad_norm": 23.68892864954692,
+      "learning_rate": 6.003305785123967e-07,
+      "logits/chosen": -0.1220703125,
+      "logits/rejected": -0.2470703125,
+      "logps/chosen": -348.0,
+      "logps/rejected": -284.0,
+      "loss": 0.1789,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.375,
+      "step": 629
+    },
+    {
+      "epoch": 0.7730061349693251,
+      "grad_norm": 33.10208379218177,
+      "learning_rate": 6e-07,
+      "logits/chosen": -0.060546875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2663,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.0625,
+      "step": 630
+    },
+    {
+      "epoch": 0.7742331288343558,
+      "grad_norm": 36.528138511875646,
+      "learning_rate": 5.996694214876033e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3321,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.875,
+      "step": 631
+    },
+    {
+      "epoch": 0.7754601226993865,
+      "grad_norm": 36.37506860284647,
+      "learning_rate": 5.993388429752066e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3056,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.34375,
+      "step": 632
+    },
+    {
+      "epoch": 0.7766871165644171,
+      "grad_norm": 40.1484232566858,
+      "learning_rate": 5.990082644628099e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.21484375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3418,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.84375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.65625,
+      "step": 633
+    },
+    {
+      "epoch": 0.7779141104294478,
+      "grad_norm": 32.83617762178678,
+      "learning_rate": 5.986776859504132e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2202,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 2.734375,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -5.1875,
+      "step": 634
+    },
+    {
+      "epoch": 0.7791411042944786,
+      "grad_norm": 29.099912658258905,
+      "learning_rate": 5.983471074380165e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2119,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.84375,
+      "step": 635
+    },
+    {
+      "epoch": 0.7803680981595092,
+      "grad_norm": 39.54581126371982,
+      "learning_rate": 5.980165289256198e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -366.0,
+      "loss": 0.3099,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -4.90625,
+      "step": 636
+    },
+    {
+      "epoch": 0.7815950920245399,
+      "grad_norm": 45.109718146173684,
+      "learning_rate": 5.976859504132231e-07,
+      "logits/chosen": -0.03125,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3522,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.0,
+      "step": 637
+    },
+    {
+      "epoch": 0.7828220858895706,
+      "grad_norm": 29.717665569657886,
+      "learning_rate": 5.973553719008264e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2213,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.9375,
+      "step": 638
+    },
+    {
+      "epoch": 0.7840490797546013,
+      "grad_norm": 38.75212680036792,
+      "learning_rate": 5.970247933884297e-07,
+      "logits/chosen": -0.048828125,
+      "logits/rejected": -0.2265625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3316,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.6484375,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.34375,
+      "step": 639
+    },
+    {
+      "epoch": 0.7852760736196319,
+      "grad_norm": 39.02043679792906,
+      "learning_rate": 5.96694214876033e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -440.0,
+      "logps/rejected": -382.0,
+      "loss": 0.2894,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -4.71875,
+      "step": 640
+    },
+    {
+      "epoch": 0.7865030674846626,
+      "grad_norm": 44.494628054471185,
+      "learning_rate": 5.963636363636364e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -348.0,
+      "loss": 0.4553,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.03125,
+      "step": 641
+    },
+    {
+      "epoch": 0.7877300613496933,
+      "grad_norm": 27.200186618621586,
+      "learning_rate": 5.960330578512397e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2487,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.5,
+      "step": 642
+    },
+    {
+      "epoch": 0.7889570552147239,
+      "grad_norm": 36.983988342062,
+      "learning_rate": 5.95702479338843e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -372.0,
+      "loss": 0.2631,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -4.53125,
+      "step": 643
+    },
+    {
+      "epoch": 0.7901840490797546,
+      "grad_norm": 39.431613095622076,
+      "learning_rate": 5.953719008264462e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2856,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.09375,
+      "step": 644
+    },
+    {
+      "epoch": 0.7914110429447853,
+      "grad_norm": 33.34429063414653,
+      "learning_rate": 5.950413223140495e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -328.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2445,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -4.9375,
+      "step": 645
+    },
+    {
+      "epoch": 0.7926380368098159,
+      "grad_norm": 34.57872089054598,
+      "learning_rate": 5.947107438016528e-07,
+      "logits/chosen": -0.0859375,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2881,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.75,
+      "step": 646
+    },
+    {
+      "epoch": 0.7938650306748466,
+      "grad_norm": 34.439204602754096,
+      "learning_rate": 5.943801652892562e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2819,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -5.125,
+      "step": 647
+    },
+    {
+      "epoch": 0.7950920245398773,
+      "grad_norm": 41.50373217977521,
+      "learning_rate": 5.940495867768595e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -324.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3682,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.859375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.84375,
+      "step": 648
+    },
+    {
+      "epoch": 0.7963190184049079,
+      "grad_norm": 43.045318488956525,
+      "learning_rate": 5.937190082644628e-07,
+      "logits/chosen": -0.0654296875,
+      "logits/rejected": -0.1650390625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.491,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.21875,
+      "step": 649
+    },
+    {
+      "epoch": 0.7975460122699386,
+      "grad_norm": 41.49765653672477,
+      "learning_rate": 5.933884297520661e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3786,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.09375,
+      "step": 650
+    },
+    {
+      "epoch": 0.7987730061349694,
+      "grad_norm": 38.92619113627064,
+      "learning_rate": 5.930578512396694e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2788,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -5.5625,
+      "step": 651
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 40.314158512080596,
+      "learning_rate": 5.927272727272728e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3015,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -5.09375,
+      "step": 652
+    },
+    {
+      "epoch": 0.8012269938650307,
+      "grad_norm": 41.91917001542575,
+      "learning_rate": 5.92396694214876e-07,
+      "logits/chosen": -0.103515625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2786,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.625,
+      "step": 653
+    },
+    {
+      "epoch": 0.8024539877300614,
+      "grad_norm": 32.50826629622854,
+      "learning_rate": 5.920661157024793e-07,
+      "logits/chosen": -0.083984375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2325,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.34375,
+      "step": 654
+    },
+    {
+      "epoch": 0.803680981595092,
+      "grad_norm": 36.66838919465664,
+      "learning_rate": 5.917355371900826e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2743,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.15625,
+      "step": 655
+    },
+    {
+      "epoch": 0.8049079754601227,
+      "grad_norm": 40.353892524687666,
+      "learning_rate": 5.914049586776859e-07,
+      "logits/chosen": -0.080078125,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3542,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -5.25,
+      "step": 656
+    },
+    {
+      "epoch": 0.8061349693251534,
+      "grad_norm": 32.71653938190321,
+      "learning_rate": 5.910743801652892e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3238,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.1875,
+      "step": 657
+    },
+    {
+      "epoch": 0.807361963190184,
+      "grad_norm": 35.63490892300594,
+      "learning_rate": 5.907438016528926e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2672,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.375,
+      "step": 658
+    },
+    {
+      "epoch": 0.8085889570552147,
+      "grad_norm": 36.34676950346043,
+      "learning_rate": 5.904132231404959e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3323,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.2177734375,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -5.6875,
+      "step": 659
+    },
+    {
+      "epoch": 0.8098159509202454,
+      "grad_norm": 31.25262331885713,
+      "learning_rate": 5.900826446280992e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2661,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.4296875,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.46875,
+      "step": 660
+    },
+    {
+      "epoch": 0.811042944785276,
+      "grad_norm": 44.04724360514248,
+      "learning_rate": 5.897520661157025e-07,
+      "logits/chosen": -0.11181640625,
+      "logits/rejected": -0.2236328125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3716,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.3125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.46875,
+      "step": 661
+    },
+    {
+      "epoch": 0.8122699386503067,
+      "grad_norm": 28.306859195207647,
+      "learning_rate": 5.894214876033057e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2117,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -5.71875,
+      "step": 662
+    },
+    {
+      "epoch": 0.8134969325153374,
+      "grad_norm": 28.755676702838585,
+      "learning_rate": 5.89090909090909e-07,
+      "logits/chosen": -0.06689453125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2615,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.09375,
+      "step": 663
+    },
+    {
+      "epoch": 0.8147239263803681,
+      "grad_norm": 30.20251070806522,
+      "learning_rate": 5.887603305785124e-07,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -358.0,
+      "loss": 0.1821,
+      "rewards/accuracies": 0.9140625,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 7.84375,
+      "rewards/rejected": -6.625,
+      "step": 664
+    },
+    {
+      "epoch": 0.8159509202453987,
+      "grad_norm": 43.59493072844384,
+      "learning_rate": 5.884297520661157e-07,
+      "logits/chosen": -0.05029296875,
+      "logits/rejected": -0.2197265625,
+      "logps/chosen": -334.0,
+      "logps/rejected": -292.0,
+      "loss": 0.2979,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.5234375,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.78125,
+      "step": 665
+    },
+    {
+      "epoch": 0.8171779141104294,
+      "grad_norm": 42.64325196007558,
+      "learning_rate": 5.88099173553719e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -348.0,
+      "loss": 0.368,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.21875,
+      "step": 666
+    },
+    {
+      "epoch": 0.8184049079754602,
+      "grad_norm": 34.27649786728366,
+      "learning_rate": 5.877685950413223e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2601,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.75,
+      "step": 667
+    },
+    {
+      "epoch": 0.8196319018404908,
+      "grad_norm": 33.27456675278737,
+      "learning_rate": 5.874380165289256e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2841,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.40625,
+      "step": 668
+    },
+    {
+      "epoch": 0.8208588957055215,
+      "grad_norm": 41.4043511264187,
+      "learning_rate": 5.871074380165289e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3017,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.375,
+      "step": 669
+    },
+    {
+      "epoch": 0.8220858895705522,
+      "grad_norm": 31.799223551777082,
+      "learning_rate": 5.867768595041323e-07,
+      "logits/chosen": -0.0191650390625,
+      "logits/rejected": -0.1650390625,
+      "logps/chosen": -324.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2541,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.125,
+      "step": 670
+    },
+    {
+      "epoch": 0.8233128834355828,
+      "grad_norm": 31.913109572694793,
+      "learning_rate": 5.864462809917355e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2407,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.0625,
+      "step": 671
+    },
+    {
+      "epoch": 0.8245398773006135,
+      "grad_norm": 32.62700830171627,
+      "learning_rate": 5.861157024793388e-07,
+      "logits/chosen": -0.109375,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -444.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2681,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.078125,
+      "rewards/margins": 7.25,
+      "rewards/rejected": -5.15625,
+      "step": 672
+    },
+    {
+      "epoch": 0.8257668711656442,
+      "grad_norm": 36.66995682138012,
+      "learning_rate": 5.857851239669421e-07,
+      "logits/chosen": -0.0498046875,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2592,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.625,
+      "step": 673
+    },
+    {
+      "epoch": 0.8269938650306748,
+      "grad_norm": 34.606200421728126,
+      "learning_rate": 5.854545454545454e-07,
+      "logits/chosen": -0.01300048828125,
+      "logits/rejected": -0.177734375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2822,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.09375,
+      "step": 674
+    },
+    {
+      "epoch": 0.8282208588957055,
+      "grad_norm": 35.038586845163366,
+      "learning_rate": 5.851239669421487e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3168,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.28125,
+      "step": 675
+    },
+    {
+      "epoch": 0.8294478527607362,
+      "grad_norm": 35.465018732105385,
+      "learning_rate": 5.847933884297521e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2534,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.71875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.53125,
+      "step": 676
+    },
+    {
+      "epoch": 0.8306748466257668,
+      "grad_norm": 36.537944400776084,
+      "learning_rate": 5.844628099173554e-07,
+      "logits/chosen": -0.06982421875,
+      "logits/rejected": -0.2451171875,
+      "logps/chosen": -334.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2916,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.84375,
+      "step": 677
+    },
+    {
+      "epoch": 0.8319018404907975,
+      "grad_norm": 33.168213995988694,
+      "learning_rate": 5.841322314049587e-07,
+      "logits/chosen": -0.1005859375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2768,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.25,
+      "step": 678
+    },
+    {
+      "epoch": 0.8331288343558282,
+      "grad_norm": 37.74716831610831,
+      "learning_rate": 5.838016528925619e-07,
+      "logits/chosen": -0.01220703125,
+      "logits/rejected": -0.1259765625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3364,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.8125,
+      "step": 679
+    },
+    {
+      "epoch": 0.8343558282208589,
+      "grad_norm": 34.79631984317766,
+      "learning_rate": 5.834710743801652e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -454.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2698,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -5.65625,
+      "step": 680
+    },
+    {
+      "epoch": 0.8355828220858895,
+      "grad_norm": 27.30848566312445,
+      "learning_rate": 5.831404958677685e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -324.0,
+      "loss": 0.1854,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -6.25,
+      "step": 681
+    },
+    {
+      "epoch": 0.8368098159509203,
+      "grad_norm": 36.85289178579729,
+      "learning_rate": 5.828099173553719e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3333,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.81640625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.65625,
+      "step": 682
+    },
+    {
+      "epoch": 0.838036809815951,
+      "grad_norm": 33.68839599065418,
+      "learning_rate": 5.824793388429752e-07,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2767,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.203125,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.8125,
+      "step": 683
+    },
+    {
+      "epoch": 0.8392638036809816,
+      "grad_norm": 35.217179760941754,
+      "learning_rate": 5.821487603305785e-07,
+      "logits/chosen": 0.000904083251953125,
+      "logits/rejected": -0.1181640625,
+      "logps/chosen": -314.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3304,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": -0.0341796875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -5.4375,
+      "step": 684
+    },
+    {
+      "epoch": 0.8404907975460123,
+      "grad_norm": 34.3907812005455,
+      "learning_rate": 5.818181818181818e-07,
+      "logits/chosen": -0.11865234375,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2858,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.9375,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -6.59375,
+      "step": 685
+    },
+    {
+      "epoch": 0.841717791411043,
+      "grad_norm": 29.480838482306872,
+      "learning_rate": 5.814876033057851e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2598,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.8125,
+      "step": 686
+    },
+    {
+      "epoch": 0.8429447852760736,
+      "grad_norm": 34.28684664510481,
+      "learning_rate": 5.811570247933884e-07,
+      "logits/chosen": -0.11767578125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2709,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.65625,
+      "step": 687
+    },
+    {
+      "epoch": 0.8441717791411043,
+      "grad_norm": 35.903679070126266,
+      "learning_rate": 5.808264462809917e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2748,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.07958984375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -6.125,
+      "step": 688
+    },
+    {
+      "epoch": 0.845398773006135,
+      "grad_norm": 37.11450186775036,
+      "learning_rate": 5.80495867768595e-07,
+      "logits/chosen": -0.12353515625,
+      "logits/rejected": -0.2197265625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3107,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.2109375,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.375,
+      "step": 689
+    },
+    {
+      "epoch": 0.8466257668711656,
+      "grad_norm": 49.86231404884124,
+      "learning_rate": 5.801652892561983e-07,
+      "logits/chosen": 0.004058837890625,
+      "logits/rejected": -0.1943359375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -350.0,
+      "loss": 0.4457,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.3125,
+      "step": 690
+    },
+    {
+      "epoch": 0.8478527607361963,
+      "grad_norm": 30.320356705626047,
+      "learning_rate": 5.798347107438016e-07,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2435,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.71875,
+      "step": 691
+    },
+    {
+      "epoch": 0.849079754601227,
+      "grad_norm": 43.95989415122109,
+      "learning_rate": 5.795041322314049e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -336.0,
+      "loss": 0.282,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -5.34375,
+      "step": 692
+    },
+    {
+      "epoch": 0.8503067484662576,
+      "grad_norm": 33.0652784939274,
+      "learning_rate": 5.791735537190082e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -442.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2934,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.203125,
+      "rewards/margins": 7.8125,
+      "rewards/rejected": -5.625,
+      "step": 693
+    },
+    {
+      "epoch": 0.8515337423312883,
+      "grad_norm": 46.044690743816275,
+      "learning_rate": 5.788429752066116e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -330.0,
+      "loss": 0.4005,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.5625,
+      "step": 694
+    },
+    {
+      "epoch": 0.852760736196319,
+      "grad_norm": 47.292538608776496,
+      "learning_rate": 5.785123966942149e-07,
+      "logits/chosen": -0.11376953125,
+      "logits/rejected": -0.208984375,
+      "logps/chosen": -328.0,
+      "logps/rejected": -302.0,
+      "loss": 0.4681,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 4.625,
+      "rewards/rejected": -3.90625,
+      "step": 695
+    },
+    {
+      "epoch": 0.8539877300613496,
+      "grad_norm": 38.43478974924762,
+      "learning_rate": 5.781818181818182e-07,
+      "logits/chosen": -0.07568359375,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3104,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.6875,
+      "step": 696
+    },
+    {
+      "epoch": 0.8552147239263803,
+      "grad_norm": 58.82691398412282,
+      "learning_rate": 5.778512396694214e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3128,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.140625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -4.78125,
+      "step": 697
+    },
+    {
+      "epoch": 0.8564417177914111,
+      "grad_norm": 37.608292873182215,
+      "learning_rate": 5.775206611570247e-07,
+      "logits/chosen": -0.080078125,
+      "logits/rejected": -0.216796875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2946,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -4.84375,
+      "step": 698
+    },
+    {
+      "epoch": 0.8576687116564418,
+      "grad_norm": 41.848029972370156,
+      "learning_rate": 5.77190082644628e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.2421875,
+      "logps/chosen": -334.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2843,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.91015625,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.71875,
+      "step": 699
+    },
+    {
+      "epoch": 0.8588957055214724,
+      "grad_norm": 39.920137771888676,
+      "learning_rate": 5.768595041322314e-07,
+      "logits/chosen": -0.054931640625,
+      "logits/rejected": -0.1640625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3259,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.34375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.25,
+      "step": 700
+    },
+    {
+      "epoch": 0.8601226993865031,
+      "grad_norm": 41.52681124295287,
+      "learning_rate": 5.765289256198347e-07,
+      "logits/chosen": -0.025146484375,
+      "logits/rejected": -0.138671875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -322.0,
+      "loss": 0.385,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.78125,
+      "step": 701
+    },
+    {
+      "epoch": 0.8613496932515338,
+      "grad_norm": 28.252991079387233,
+      "learning_rate": 5.76198347107438e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -330.0,
+      "loss": 0.1786,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -5.21875,
+      "step": 702
+    },
+    {
+      "epoch": 0.8625766871165644,
+      "grad_norm": 31.56795802217109,
+      "learning_rate": 5.758677685950413e-07,
+      "logits/chosen": -0.046142578125,
+      "logits/rejected": -0.1884765625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2909,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.3125,
+      "step": 703
+    },
+    {
+      "epoch": 0.8638036809815951,
+      "grad_norm": 34.11635728276886,
+      "learning_rate": 5.755371900826446e-07,
+      "logits/chosen": -0.1201171875,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -320.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3105,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 6.1875,
+      "rewards/rejected": -4.9375,
+      "step": 704
+    },
+    {
+      "epoch": 0.8650306748466258,
+      "grad_norm": 38.95077133588588,
+      "learning_rate": 5.75206611570248e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2767,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.265625,
+      "rewards/margins": 7.15625,
+      "rewards/rejected": -4.875,
+      "step": 705
+    },
+    {
+      "epoch": 0.8662576687116564,
+      "grad_norm": 35.267737415428456,
+      "learning_rate": 5.748760330578512e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3068,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.53125,
+      "step": 706
+    },
+    {
+      "epoch": 0.8674846625766871,
+      "grad_norm": 34.32550780952695,
+      "learning_rate": 5.745454545454545e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2683,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.25,
+      "step": 707
+    },
+    {
+      "epoch": 0.8687116564417178,
+      "grad_norm": 31.650191661242673,
+      "learning_rate": 5.742148760330578e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2379,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -4.875,
+      "step": 708
+    },
+    {
+      "epoch": 0.8699386503067484,
+      "grad_norm": 25.32301991794396,
+      "learning_rate": 5.738842975206611e-07,
+      "logits/chosen": -0.06640625,
+      "logits/rejected": -0.23046875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -340.0,
+      "loss": 0.185,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.9765625,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -5.53125,
+      "step": 709
+    },
+    {
+      "epoch": 0.8711656441717791,
+      "grad_norm": 33.22472192244502,
+      "learning_rate": 5.735537190082644e-07,
+      "logits/chosen": -0.017333984375,
+      "logits/rejected": -0.15234375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -312.0,
+      "loss": 0.299,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.625,
+      "step": 710
+    },
+    {
+      "epoch": 0.8723926380368098,
+      "grad_norm": 29.02267964751041,
+      "learning_rate": 5.732231404958678e-07,
+      "logits/chosen": -0.042724609375,
+      "logits/rejected": -0.236328125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2551,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.78125,
+      "step": 711
+    },
+    {
+      "epoch": 0.8736196319018404,
+      "grad_norm": 40.668984053435686,
+      "learning_rate": 5.728925619834711e-07,
+      "logits/chosen": -0.03857421875,
+      "logits/rejected": -0.1923828125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3587,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.375,
+      "step": 712
+    },
+    {
+      "epoch": 0.8748466257668711,
+      "grad_norm": 33.08084767770349,
+      "learning_rate": 5.725619834710744e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2665,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.15625,
+      "step": 713
+    },
+    {
+      "epoch": 0.8760736196319019,
+      "grad_norm": 36.45936464644564,
+      "learning_rate": 5.722314049586777e-07,
+      "logits/chosen": -0.0159912109375,
+      "logits/rejected": -0.2333984375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2938,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.75,
+      "step": 714
+    },
+    {
+      "epoch": 0.8773006134969326,
+      "grad_norm": 30.913206903477224,
+      "learning_rate": 5.719008264462809e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2684,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.53125,
+      "step": 715
+    },
+    {
+      "epoch": 0.8785276073619632,
+      "grad_norm": 41.616080953336535,
+      "learning_rate": 5.715702479338842e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3142,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.625,
+      "step": 716
+    },
+    {
+      "epoch": 0.8797546012269939,
+      "grad_norm": 34.18654539125182,
+      "learning_rate": 5.712396694214876e-07,
+      "logits/chosen": -0.09423828125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -276.0,
+      "loss": 0.2911,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.9296875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.3125,
+      "step": 717
+    },
+    {
+      "epoch": 0.8809815950920246,
+      "grad_norm": 28.09115304120503,
+      "learning_rate": 5.709090909090909e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2254,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.859375,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.0625,
+      "step": 718
+    },
+    {
+      "epoch": 0.8822085889570552,
+      "grad_norm": 41.08426714668054,
+      "learning_rate": 5.705785123966942e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3011,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.25,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -4.9375,
+      "step": 719
+    },
+    {
+      "epoch": 0.8834355828220859,
+      "grad_norm": 28.518925312083375,
+      "learning_rate": 5.702479338842975e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2223,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.265625,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -4.78125,
+      "step": 720
+    },
+    {
+      "epoch": 0.8846625766871166,
+      "grad_norm": 26.149570167864507,
+      "learning_rate": 5.699173553719008e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2073,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.890625,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -6.09375,
+      "step": 721
+    },
+    {
+      "epoch": 0.8858895705521472,
+      "grad_norm": 30.929567326750046,
+      "learning_rate": 5.695867768595041e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2229,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.859375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.125,
+      "step": 722
+    },
+    {
+      "epoch": 0.8871165644171779,
+      "grad_norm": 34.37293415398814,
+      "learning_rate": 5.692561983471075e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2808,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -5.4375,
+      "step": 723
+    },
+    {
+      "epoch": 0.8883435582822086,
+      "grad_norm": 33.90677929130864,
+      "learning_rate": 5.689256198347107e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -308.0,
+      "loss": 0.254,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.7265625,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.21875,
+      "step": 724
+    },
+    {
+      "epoch": 0.8895705521472392,
+      "grad_norm": 34.97307578381243,
+      "learning_rate": 5.68595041322314e-07,
+      "logits/chosen": -0.055419921875,
+      "logits/rejected": -0.2314453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3044,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.15625,
+      "step": 725
+    },
+    {
+      "epoch": 0.8907975460122699,
+      "grad_norm": 28.627835648582934,
+      "learning_rate": 5.682644628099173e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -336.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2135,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.9765625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.5625,
+      "step": 726
+    },
+    {
+      "epoch": 0.8920245398773006,
+      "grad_norm": 34.81272514902735,
+      "learning_rate": 5.679338842975206e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -324.0,
+      "loss": 0.248,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.15625,
+      "step": 727
+    },
+    {
+      "epoch": 0.8932515337423312,
+      "grad_norm": 37.8936877345969,
+      "learning_rate": 5.676033057851239e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -398.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2946,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 7.65625,
+      "rewards/rejected": -5.625,
+      "step": 728
+    },
+    {
+      "epoch": 0.894478527607362,
+      "grad_norm": 29.90567314985844,
+      "learning_rate": 5.672727272727273e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2088,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 7.6875,
+      "rewards/rejected": -6.0625,
+      "step": 729
+    },
+    {
+      "epoch": 0.8957055214723927,
+      "grad_norm": 31.408249750098918,
+      "learning_rate": 5.669421487603306e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -398.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2204,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -5.75,
+      "step": 730
+    },
+    {
+      "epoch": 0.8969325153374234,
+      "grad_norm": 33.91624972638253,
+      "learning_rate": 5.666115702479339e-07,
+      "logits/chosen": -0.052734375,
+      "logits/rejected": -0.1845703125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2736,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.8125,
+      "step": 731
+    },
+    {
+      "epoch": 0.898159509202454,
+      "grad_norm": 40.7773464371931,
+      "learning_rate": 5.662809917355372e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3647,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.375,
+      "step": 732
+    },
+    {
+      "epoch": 0.8993865030674847,
+      "grad_norm": 31.001224216873553,
+      "learning_rate": 5.659504132231404e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2212,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.84375,
+      "step": 733
+    },
+    {
+      "epoch": 0.9006134969325154,
+      "grad_norm": 37.803749225628884,
+      "learning_rate": 5.656198347107437e-07,
+      "logits/chosen": -0.08837890625,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2954,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 7.5,
+      "rewards/rejected": -6.25,
+      "step": 734
+    },
+    {
+      "epoch": 0.901840490797546,
+      "grad_norm": 30.18047116909555,
+      "learning_rate": 5.652892561983471e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -372.0,
+      "loss": 0.1824,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -6.8125,
+      "step": 735
+    },
+    {
+      "epoch": 0.9030674846625767,
+      "grad_norm": 38.572832638835585,
+      "learning_rate": 5.649586776859504e-07,
+      "logits/chosen": -0.025634765625,
+      "logits/rejected": -0.1552734375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3608,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.0908203125,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -6.3125,
+      "step": 736
+    },
+    {
+      "epoch": 0.9042944785276074,
+      "grad_norm": 46.258369054797996,
+      "learning_rate": 5.646280991735537e-07,
+      "logits/chosen": -0.12451171875,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4383,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.1826171875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -6.25,
+      "step": 737
+    },
+    {
+      "epoch": 0.905521472392638,
+      "grad_norm": 32.39664225178629,
+      "learning_rate": 5.64297520661157e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -442.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2588,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 7.5625,
+      "rewards/rejected": -6.28125,
+      "step": 738
+    },
+    {
+      "epoch": 0.9067484662576687,
+      "grad_norm": 34.18126494383032,
+      "learning_rate": 5.639669421487603e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -344.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2836,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.2470703125,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -6.15625,
+      "step": 739
+    },
+    {
+      "epoch": 0.9079754601226994,
+      "grad_norm": 35.15083279583394,
+      "learning_rate": 5.636363636363636e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -342.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3072,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.2431640625,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -6.6875,
+      "step": 740
+    },
+    {
+      "epoch": 0.90920245398773,
+      "grad_norm": 33.535600643328806,
+      "learning_rate": 5.63305785123967e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2851,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -6.8125,
+      "step": 741
+    },
+    {
+      "epoch": 0.9104294478527607,
+      "grad_norm": 31.52809779595991,
+      "learning_rate": 5.629752066115702e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2244,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.90625,
+      "step": 742
+    },
+    {
+      "epoch": 0.9116564417177914,
+      "grad_norm": 35.37117075886455,
+      "learning_rate": 5.626446280991735e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -380.0,
+      "loss": 0.2769,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.8125,
+      "step": 743
+    },
+    {
+      "epoch": 0.912883435582822,
+      "grad_norm": 26.42035749866701,
+      "learning_rate": 5.623140495867768e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -330.0,
+      "loss": 0.1917,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.75,
+      "step": 744
+    },
+    {
+      "epoch": 0.9141104294478528,
+      "grad_norm": 35.853268232755404,
+      "learning_rate": 5.619834710743801e-07,
+      "logits/chosen": -0.08251953125,
+      "logits/rejected": -0.1865234375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3009,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.390625,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.875,
+      "step": 745
+    },
+    {
+      "epoch": 0.9153374233128835,
+      "grad_norm": 43.35686834196019,
+      "learning_rate": 5.616528925619834e-07,
+      "logits/chosen": -0.0849609375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -346.0,
+      "loss": 0.317,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -6.0,
+      "step": 746
+    },
+    {
+      "epoch": 0.9165644171779141,
+      "grad_norm": 34.17444251605582,
+      "learning_rate": 5.613223140495868e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2811,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.41796875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.78125,
+      "step": 747
+    },
+    {
+      "epoch": 0.9177914110429448,
+      "grad_norm": 40.3104413419788,
+      "learning_rate": 5.609917355371901e-07,
+      "logits/chosen": -0.05224609375,
+      "logits/rejected": -0.169921875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3882,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.0279541015625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -5.625,
+      "step": 748
+    },
+    {
+      "epoch": 0.9190184049079755,
+      "grad_norm": 40.57709610969512,
+      "learning_rate": 5.606611570247934e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3642,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.98828125,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -6.09375,
+      "step": 749
+    },
+    {
+      "epoch": 0.9202453987730062,
+      "grad_norm": 43.682412563973024,
+      "learning_rate": 5.603305785123967e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3704,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.63671875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.875,
+      "step": 750
+    },
+    {
+      "epoch": 0.9214723926380368,
+      "grad_norm": 42.01608666567515,
+      "learning_rate": 5.599999999999999e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3006,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 7.4375,
+      "rewards/rejected": -6.28125,
+      "step": 751
+    },
+    {
+      "epoch": 0.9226993865030675,
+      "grad_norm": 32.96496448782055,
+      "learning_rate": 5.596694214876033e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2435,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 7.75,
+      "rewards/rejected": -6.34375,
+      "step": 752
+    },
+    {
+      "epoch": 0.9239263803680982,
+      "grad_norm": 35.77444345136624,
+      "learning_rate": 5.593388429752066e-07,
+      "logits/chosen": -0.048095703125,
+      "logits/rejected": -0.1923828125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2851,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.796875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.5625,
+      "step": 753
+    },
+    {
+      "epoch": 0.9251533742331288,
+      "grad_norm": 39.56695959824003,
+      "learning_rate": 5.590082644628099e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3588,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.59375,
+      "step": 754
+    },
+    {
+      "epoch": 0.9263803680981595,
+      "grad_norm": 33.47777157003214,
+      "learning_rate": 5.586776859504132e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2421,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.6875,
+      "step": 755
+    },
+    {
+      "epoch": 0.9276073619631902,
+      "grad_norm": 30.73963254337421,
+      "learning_rate": 5.583471074380165e-07,
+      "logits/chosen": -0.0732421875,
+      "logits/rejected": -0.1669921875,
+      "logps/chosen": -332.0,
+      "logps/rejected": -308.0,
+      "loss": 0.248,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.322265625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.75,
+      "step": 756
+    },
+    {
+      "epoch": 0.9288343558282208,
+      "grad_norm": 25.863947966401952,
+      "learning_rate": 5.580165289256198e-07,
+      "logits/chosen": -0.0252685546875,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2117,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.375,
+      "step": 757
+    },
+    {
+      "epoch": 0.9300613496932515,
+      "grad_norm": 40.231833773135115,
+      "learning_rate": 5.576859504132232e-07,
+      "logits/chosen": -0.08837890625,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3497,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.46875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.59375,
+      "step": 758
+    },
+    {
+      "epoch": 0.9312883435582822,
+      "grad_norm": 34.14796678847348,
+      "learning_rate": 5.573553719008265e-07,
+      "logits/chosen": -0.0595703125,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2511,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.34375,
+      "step": 759
+    },
+    {
+      "epoch": 0.9325153374233128,
+      "grad_norm": 29.867916639384045,
+      "learning_rate": 5.570247933884297e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2724,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 7.8125,
+      "rewards/rejected": -5.71875,
+      "step": 760
+    },
+    {
+      "epoch": 0.9337423312883436,
+      "grad_norm": 38.414301746929866,
+      "learning_rate": 5.56694214876033e-07,
+      "logits/chosen": -0.0264892578125,
+      "logits/rejected": -0.1669921875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2874,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.375,
+      "step": 761
+    },
+    {
+      "epoch": 0.9349693251533743,
+      "grad_norm": 40.447076836962,
+      "learning_rate": 5.563636363636363e-07,
+      "logits/chosen": -0.01397705078125,
+      "logits/rejected": -0.19921875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2888,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.6875,
+      "step": 762
+    },
+    {
+      "epoch": 0.9361963190184049,
+      "grad_norm": 31.879754151772115,
+      "learning_rate": 5.560330578512396e-07,
+      "logits/chosen": -0.07373046875,
+      "logits/rejected": -0.171875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2711,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.4375,
+      "step": 763
+    },
+    {
+      "epoch": 0.9374233128834356,
+      "grad_norm": 35.145542304507835,
+      "learning_rate": 5.55702479338843e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -322.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2896,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.26171875,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -6.0,
+      "step": 764
+    },
+    {
+      "epoch": 0.9386503067484663,
+      "grad_norm": 40.67019743057879,
+      "learning_rate": 5.553719008264463e-07,
+      "logits/chosen": -0.0595703125,
+      "logits/rejected": -0.19140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3457,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.76171875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.96875,
+      "step": 765
+    },
+    {
+      "epoch": 0.939877300613497,
+      "grad_norm": 32.24931969694431,
+      "learning_rate": 5.550413223140496e-07,
+      "logits/chosen": -0.11767578125,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -326.0,
+      "loss": 0.274,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.5625,
+      "step": 766
+    },
+    {
+      "epoch": 0.9411042944785276,
+      "grad_norm": 34.46831261260643,
+      "learning_rate": 5.547107438016529e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3035,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.953125,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.9375,
+      "step": 767
+    },
+    {
+      "epoch": 0.9423312883435583,
+      "grad_norm": 36.15353177119284,
+      "learning_rate": 5.543801652892561e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -372.0,
+      "loss": 0.316,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -5.75,
+      "step": 768
+    },
+    {
+      "epoch": 0.943558282208589,
+      "grad_norm": 29.41748843520645,
+      "learning_rate": 5.540495867768594e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2432,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.6328125,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -6.375,
+      "step": 769
+    },
+    {
+      "epoch": 0.9447852760736196,
+      "grad_norm": 41.73128837710126,
+      "learning_rate": 5.537190082644628e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3004,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -5.46875,
+      "step": 770
+    },
+    {
+      "epoch": 0.9460122699386503,
+      "grad_norm": 42.981692808778455,
+      "learning_rate": 5.533884297520661e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.201171875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -346.0,
+      "loss": 0.429,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 0.0869140625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -5.5625,
+      "step": 771
+    },
+    {
+      "epoch": 0.947239263803681,
+      "grad_norm": 35.00554577923317,
+      "learning_rate": 5.530578512396694e-07,
+      "logits/chosen": -0.038330078125,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3114,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.6875,
+      "step": 772
+    },
+    {
+      "epoch": 0.9484662576687116,
+      "grad_norm": 36.88335000648106,
+      "learning_rate": 5.527272727272727e-07,
+      "logits/chosen": -0.0361328125,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3175,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.1513671875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.78125,
+      "step": 773
+    },
+    {
+      "epoch": 0.9496932515337423,
+      "grad_norm": 38.52772066032803,
+      "learning_rate": 5.52396694214876e-07,
+      "logits/chosen": -0.058837890625,
+      "logits/rejected": -0.2001953125,
+      "logps/chosen": -352.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3665,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": -0.330078125,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -5.0625,
+      "step": 774
+    },
+    {
+      "epoch": 0.950920245398773,
+      "grad_norm": 38.30272199053397,
+      "learning_rate": 5.520661157024793e-07,
+      "logits/chosen": 0.0306396484375,
+      "logits/rejected": -0.162109375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3245,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.03125,
+      "step": 775
+    },
+    {
+      "epoch": 0.9521472392638037,
+      "grad_norm": 33.16873912890908,
+      "learning_rate": 5.517355371900827e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3113,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -5.375,
+      "step": 776
+    },
+    {
+      "epoch": 0.9533742331288344,
+      "grad_norm": 35.432794054015964,
+      "learning_rate": 5.514049586776859e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3098,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.8125,
+      "step": 777
+    },
+    {
+      "epoch": 0.9546012269938651,
+      "grad_norm": 35.189286152943,
+      "learning_rate": 5.510743801652892e-07,
+      "logits/chosen": -0.1083984375,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -314.0,
+      "loss": 0.251,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.443359375,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -5.34375,
+      "step": 778
+    },
+    {
+      "epoch": 0.9558282208588957,
+      "grad_norm": 41.96194349470727,
+      "learning_rate": 5.507438016528925e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2827,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.96875,
+      "step": 779
+    },
+    {
+      "epoch": 0.9570552147239264,
+      "grad_norm": 35.73940962793884,
+      "learning_rate": 5.504132231404958e-07,
+      "logits/chosen": -0.059326171875,
+      "logits/rejected": -0.2353515625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3415,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.2138671875,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -5.4375,
+      "step": 780
+    },
+    {
+      "epoch": 0.9582822085889571,
+      "grad_norm": 26.50283260922046,
+      "learning_rate": 5.500826446280991e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -342.0,
+      "loss": 0.1906,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.4375,
+      "step": 781
+    },
+    {
+      "epoch": 0.9595092024539877,
+      "grad_norm": 29.07388144480791,
+      "learning_rate": 5.497520661157025e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2204,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.5625,
+      "step": 782
+    },
+    {
+      "epoch": 0.9607361963190184,
+      "grad_norm": 36.281519570454215,
+      "learning_rate": 5.494214876033058e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2544,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -6.09375,
+      "step": 783
+    },
+    {
+      "epoch": 0.9619631901840491,
+      "grad_norm": 41.7797714543972,
+      "learning_rate": 5.490909090909091e-07,
+      "logits/chosen": -0.11181640625,
+      "logits/rejected": -0.162109375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3585,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.1767578125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.625,
+      "step": 784
+    },
+    {
+      "epoch": 0.9631901840490797,
+      "grad_norm": 31.81554374340158,
+      "learning_rate": 5.487603305785124e-07,
+      "logits/chosen": -0.0751953125,
+      "logits/rejected": -0.24609375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2472,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.88671875,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -5.8125,
+      "step": 785
+    },
+    {
+      "epoch": 0.9644171779141104,
+      "grad_norm": 27.670018917456815,
+      "learning_rate": 5.484297520661156e-07,
+      "logits/chosen": -0.0732421875,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2222,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 0.4921875,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -6.25,
+      "step": 786
+    },
+    {
+      "epoch": 0.9656441717791411,
+      "grad_norm": 37.16124889650476,
+      "learning_rate": 5.480991735537189e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -6.4375,
+      "step": 787
+    },
+    {
+      "epoch": 0.9668711656441717,
+      "grad_norm": 24.391962735004444,
+      "learning_rate": 5.477685950413223e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -352.0,
+      "loss": 0.1706,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -6.25,
+      "step": 788
+    },
+    {
+      "epoch": 0.9680981595092024,
+      "grad_norm": 32.024862187416,
+      "learning_rate": 5.474380165289256e-07,
+      "logits/chosen": -0.07958984375,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2843,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.408203125,
+      "rewards/margins": 6.1875,
+      "rewards/rejected": -5.78125,
+      "step": 789
+    },
+    {
+      "epoch": 0.9693251533742331,
+      "grad_norm": 27.51754897072937,
+      "learning_rate": 5.471074380165289e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -316.0,
+      "loss": 0.1866,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -6.03125,
+      "step": 790
+    },
+    {
+      "epoch": 0.9705521472392638,
+      "grad_norm": 37.440979455018265,
+      "learning_rate": 5.467768595041322e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2983,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 7.59375,
+      "rewards/rejected": -6.375,
+      "step": 791
+    },
+    {
+      "epoch": 0.9717791411042945,
+      "grad_norm": 36.704827910830204,
+      "learning_rate": 5.464462809917355e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -430.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3431,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.03125,
+      "step": 792
+    },
+    {
+      "epoch": 0.9730061349693252,
+      "grad_norm": 37.60757704339766,
+      "learning_rate": 5.461157024793389e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3237,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.4375,
+      "step": 793
+    },
+    {
+      "epoch": 0.9742331288343559,
+      "grad_norm": 37.59792190871499,
+      "learning_rate": 5.457851239669422e-07,
+      "logits/chosen": -0.103515625,
+      "logits/rejected": -0.2431640625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3909,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.21875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -5.78125,
+      "step": 794
+    },
+    {
+      "epoch": 0.9754601226993865,
+      "grad_norm": 28.85691430451381,
+      "learning_rate": 5.454545454545454e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2394,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.90625,
+      "step": 795
+    },
+    {
+      "epoch": 0.9766871165644172,
+      "grad_norm": 41.11938718622486,
+      "learning_rate": 5.451239669421487e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2984,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": -0.0673828125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -6.375,
+      "step": 796
+    },
+    {
+      "epoch": 0.9779141104294479,
+      "grad_norm": 30.942308293626358,
+      "learning_rate": 5.44793388429752e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -308.0,
+      "loss": 0.254,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.625,
+      "step": 797
+    },
+    {
+      "epoch": 0.9791411042944785,
+      "grad_norm": 36.786308023977305,
+      "learning_rate": 5.444628099173553e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2823,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.76953125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -6.125,
+      "step": 798
+    },
+    {
+      "epoch": 0.9803680981595092,
+      "grad_norm": 40.89607573095922,
+      "learning_rate": 5.441322314049587e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3222,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.78125,
+      "step": 799
+    },
+    {
+      "epoch": 0.9815950920245399,
+      "grad_norm": 36.10519822526459,
+      "learning_rate": 5.43801652892562e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2658,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.61328125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -6.0,
+      "step": 800
+    },
+    {
+      "epoch": 0.9828220858895705,
+      "grad_norm": 42.48783397210697,
+      "learning_rate": 5.434710743801653e-07,
+      "logits/chosen": -0.17578125,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -366.0,
+      "loss": 0.3944,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.1875,
+      "step": 801
+    },
+    {
+      "epoch": 0.9840490797546012,
+      "grad_norm": 43.552763355211724,
+      "learning_rate": 5.431404958677686e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3899,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.453125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.84375,
+      "step": 802
+    },
+    {
+      "epoch": 0.9852760736196319,
+      "grad_norm": 32.09849729457811,
+      "learning_rate": 5.428099173553719e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2997,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.78125,
+      "step": 803
+    },
+    {
+      "epoch": 0.9865030674846625,
+      "grad_norm": 32.33615160143916,
+      "learning_rate": 5.424793388429751e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.240234375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3208,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.251953125,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -5.65625,
+      "step": 804
+    },
+    {
+      "epoch": 0.9877300613496932,
+      "grad_norm": 39.61512778302868,
+      "learning_rate": 5.421487603305785e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -380.0,
+      "loss": 0.3968,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.96875,
+      "step": 805
+    },
+    {
+      "epoch": 0.9889570552147239,
+      "grad_norm": 40.74080836640833,
+      "learning_rate": 5.418181818181818e-07,
+      "logits/chosen": -0.01202392578125,
+      "logits/rejected": -0.1474609375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -368.0,
+      "loss": 0.3811,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.34375,
+      "step": 806
+    },
+    {
+      "epoch": 0.9901840490797545,
+      "grad_norm": 36.652039792723684,
+      "learning_rate": 5.414876033057851e-07,
+      "logits/chosen": -0.1005859375,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2327,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 7.21875,
+      "rewards/rejected": -6.125,
+      "step": 807
+    },
+    {
+      "epoch": 0.9914110429447853,
+      "grad_norm": 33.57961890786831,
+      "learning_rate": 5.411570247933884e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3091,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.11572265625,
+      "rewards/margins": 6.65625,
+      "rewards/rejected": -6.53125,
+      "step": 808
+    },
+    {
+      "epoch": 0.992638036809816,
+      "grad_norm": 35.60132202210017,
+      "learning_rate": 5.408264462809917e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2618,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 7.875,
+      "rewards/rejected": -6.28125,
+      "step": 809
+    },
+    {
+      "epoch": 0.9938650306748467,
+      "grad_norm": 33.977380429457895,
+      "learning_rate": 5.40495867768595e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -368.0,
+      "loss": 0.227,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 7.28125,
+      "rewards/rejected": -6.0625,
+      "step": 810
+    },
+    {
+      "epoch": 0.9950920245398773,
+      "grad_norm": 31.635546850530186,
+      "learning_rate": 5.401652892561984e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2741,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.71875,
+      "step": 811
+    },
+    {
+      "epoch": 0.996319018404908,
+      "grad_norm": 40.80563506539147,
+      "learning_rate": 5.398347107438017e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2687,
+      "rewards/accuracies": 0.9140625,
+      "rewards/chosen": 0.4609375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -6.53125,
+      "step": 812
+    },
+    {
+      "epoch": 0.9975460122699387,
+      "grad_norm": 30.19996782105712,
+      "learning_rate": 5.395041322314049e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -338.0,
+      "logps/rejected": -324.0,
+      "loss": 0.1913,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.87109375,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -6.1875,
+      "step": 813
+    },
+    {
+      "epoch": 0.9987730061349693,
+      "grad_norm": 37.73036986625575,
+      "learning_rate": 5.391735537190082e-07,
+      "logits/chosen": -0.11767578125,
+      "logits/rejected": -0.201171875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3425,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": -0.1416015625,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -5.65625,
+      "step": 814
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 40.870567834511405,
+      "learning_rate": 5.388429752066115e-07,
+      "logits/chosen": -0.0458984375,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2125,
+      "rewards/accuracies": 0.8571429252624512,
+      "rewards/chosen": 0.59375,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -6.28125,
+      "step": 815
+    },
+    {
+      "epoch": 1.0012269938650307,
+      "grad_norm": 8.447732700524874,
+      "learning_rate": 5.385123966942148e-07,
+      "logits/chosen": -0.099609375,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -318.0,
+      "loss": 0.0478,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.75,
+      "step": 816
+    },
+    {
+      "epoch": 1.0024539877300613,
+      "grad_norm": 10.238235833774846,
+      "learning_rate": 5.381818181818182e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -320.0,
+      "loss": 0.0599,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.8828125,
+      "rewards/margins": 7.59375,
+      "rewards/rejected": -6.71875,
+      "step": 817
+    },
+    {
+      "epoch": 1.003680981595092,
+      "grad_norm": 8.509110400177521,
+      "learning_rate": 5.378512396694215e-07,
+      "logits/chosen": -0.248046875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0469,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -6.53125,
+      "step": 818
+    },
+    {
+      "epoch": 1.0049079754601227,
+      "grad_norm": 11.32230173464338,
+      "learning_rate": 5.375206611570248e-07,
+      "logits/chosen": -0.1083984375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0506,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -6.65625,
+      "step": 819
+    },
+    {
+      "epoch": 1.0061349693251533,
+      "grad_norm": 9.131424372223067,
+      "learning_rate": 5.371900826446281e-07,
+      "logits/chosen": -0.0751953125,
+      "logits/rejected": -0.216796875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0535,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.140625,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -5.78125,
+      "step": 820
+    },
+    {
+      "epoch": 1.007361963190184,
+      "grad_norm": 16.500866897905247,
+      "learning_rate": 5.368595041322314e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -344.0,
+      "logps/rejected": -326.0,
+      "loss": 0.083,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.0625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.4375,
+      "step": 821
+    },
+    {
+      "epoch": 1.0085889570552147,
+      "grad_norm": 10.055056560721704,
+      "learning_rate": 5.365289256198346e-07,
+      "logits/chosen": -0.1513671875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0514,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.9765625,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -6.03125,
+      "step": 822
+    },
+    {
+      "epoch": 1.0098159509202453,
+      "grad_norm": 8.750164950681329,
+      "learning_rate": 5.36198347107438e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0496,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.265625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.21875,
+      "step": 823
+    },
+    {
+      "epoch": 1.011042944785276,
+      "grad_norm": 8.95759785702542,
+      "learning_rate": 5.358677685950413e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0275,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.203125,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -5.9375,
+      "step": 824
+    },
+    {
+      "epoch": 1.0122699386503067,
+      "grad_norm": 8.774577043038853,
+      "learning_rate": 5.355371900826446e-07,
+      "logits/chosen": -0.021484375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0431,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -6.3125,
+      "step": 825
+    },
+    {
+      "epoch": 1.0134969325153373,
+      "grad_norm": 13.971484339500186,
+      "learning_rate": 5.352066115702479e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0773,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 7.375,
+      "rewards/rejected": -5.78125,
+      "step": 826
+    },
+    {
+      "epoch": 1.014723926380368,
+      "grad_norm": 9.213792271740255,
+      "learning_rate": 5.348760330578512e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -318.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.71875,
+      "step": 827
+    },
+    {
+      "epoch": 1.0159509202453987,
+      "grad_norm": 12.070002953910205,
+      "learning_rate": 5.345454545454545e-07,
+      "logits/chosen": -0.025634765625,
+      "logits/rejected": -0.220703125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -310.0,
+      "loss": 0.0745,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.7265625,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.6875,
+      "step": 828
+    },
+    {
+      "epoch": 1.0171779141104293,
+      "grad_norm": 14.6095484944048,
+      "learning_rate": 5.342148760330579e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0873,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 2.0625,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.03125,
+      "step": 829
+    },
+    {
+      "epoch": 1.01840490797546,
+      "grad_norm": 10.563003784443062,
+      "learning_rate": 5.338842975206612e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0528,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 2.578125,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -6.625,
+      "step": 830
+    },
+    {
+      "epoch": 1.019631901840491,
+      "grad_norm": 6.950594992951885,
+      "learning_rate": 5.335537190082644e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0425,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.328125,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -6.75,
+      "step": 831
+    },
+    {
+      "epoch": 1.0208588957055216,
+      "grad_norm": 18.903093292372017,
+      "learning_rate": 5.332231404958677e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0753,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -6.1875,
+      "step": 832
+    },
+    {
+      "epoch": 1.0220858895705522,
+      "grad_norm": 20.846144434558045,
+      "learning_rate": 5.32892561983471e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -442.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0754,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -6.625,
+      "step": 833
+    },
+    {
+      "epoch": 1.023312883435583,
+      "grad_norm": 8.267162808665637,
+      "learning_rate": 5.325619834710743e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.173828125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0429,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -6.25,
+      "step": 834
+    },
+    {
+      "epoch": 1.0245398773006136,
+      "grad_norm": 10.37076437566513,
+      "learning_rate": 5.322314049586777e-07,
+      "logits/chosen": -0.0859375,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0656,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -6.3125,
+      "step": 835
+    },
+    {
+      "epoch": 1.0257668711656442,
+      "grad_norm": 15.10396868081453,
+      "learning_rate": 5.31900826446281e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0572,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -6.9375,
+      "step": 836
+    },
+    {
+      "epoch": 1.026993865030675,
+      "grad_norm": 9.458003649126487,
+      "learning_rate": 5.315702479338843e-07,
+      "logits/chosen": -0.041015625,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0542,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 2.0625,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -6.75,
+      "step": 837
+    },
+    {
+      "epoch": 1.0282208588957056,
+      "grad_norm": 9.28143966396222,
+      "learning_rate": 5.312396694214876e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0463,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -7.5,
+      "step": 838
+    },
+    {
+      "epoch": 1.0294478527607362,
+      "grad_norm": 26.604715129647236,
+      "learning_rate": 5.309090909090909e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0866,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -6.84375,
+      "step": 839
+    },
+    {
+      "epoch": 1.030674846625767,
+      "grad_norm": 9.15851027720581,
+      "learning_rate": 5.305785123966941e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0476,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -7.84375,
+      "step": 840
+    },
+    {
+      "epoch": 1.0319018404907976,
+      "grad_norm": 9.875999818863187,
+      "learning_rate": 5.302479338842975e-07,
+      "logits/chosen": -0.0849609375,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0536,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.46875,
+      "step": 841
+    },
+    {
+      "epoch": 1.0331288343558283,
+      "grad_norm": 8.966946429222945,
+      "learning_rate": 5.299173553719008e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -448.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0535,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -6.96875,
+      "step": 842
+    },
+    {
+      "epoch": 1.034355828220859,
+      "grad_norm": 8.533895413196568,
+      "learning_rate": 5.295867768595041e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -450.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.0625,
+      "step": 843
+    },
+    {
+      "epoch": 1.0355828220858896,
+      "grad_norm": 8.091996676249778,
+      "learning_rate": 5.292561983471074e-07,
+      "logits/chosen": -0.07861328125,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0398,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.03125,
+      "step": 844
+    },
+    {
+      "epoch": 1.0368098159509203,
+      "grad_norm": 12.849424838058209,
+      "learning_rate": 5.289256198347107e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0503,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -6.28125,
+      "step": 845
+    },
+    {
+      "epoch": 1.038036809815951,
+      "grad_norm": 8.52014381356995,
+      "learning_rate": 5.285950413223141e-07,
+      "logits/chosen": -0.0546875,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0457,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.1875,
+      "step": 846
+    },
+    {
+      "epoch": 1.0392638036809816,
+      "grad_norm": 9.894794847731864,
+      "learning_rate": 5.282644628099174e-07,
+      "logits/chosen": -0.07861328125,
+      "logits/rejected": -0.2353515625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.6875,
+      "step": 847
+    },
+    {
+      "epoch": 1.0404907975460123,
+      "grad_norm": 7.67098466898408,
+      "learning_rate": 5.279338842975206e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0467,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.8515625,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.5625,
+      "step": 848
+    },
+    {
+      "epoch": 1.041717791411043,
+      "grad_norm": 13.659276587750096,
+      "learning_rate": 5.276033057851239e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.2080078125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0674,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.0,
+      "step": 849
+    },
+    {
+      "epoch": 1.0429447852760736,
+      "grad_norm": 8.672069845350178,
+      "learning_rate": 5.272727272727272e-07,
+      "logits/chosen": -0.1513671875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -342.0,
+      "loss": 0.032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.34375,
+      "step": 850
+    },
+    {
+      "epoch": 1.0441717791411043,
+      "grad_norm": 4.966708191691141,
+      "learning_rate": 5.269421487603305e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0252,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -7.5625,
+      "step": 851
+    },
+    {
+      "epoch": 1.045398773006135,
+      "grad_norm": 7.342481591915201,
+      "learning_rate": 5.266115702479339e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0379,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.828125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -6.90625,
+      "step": 852
+    },
+    {
+      "epoch": 1.0466257668711656,
+      "grad_norm": 8.819410735961545,
+      "learning_rate": 5.262809917355372e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.9375,
+      "step": 853
+    },
+    {
+      "epoch": 1.0478527607361963,
+      "grad_norm": 13.464418269579307,
+      "learning_rate": 5.259504132231405e-07,
+      "logits/chosen": -0.0966796875,
+      "logits/rejected": -0.232421875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0633,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 7.65625,
+      "rewards/rejected": -6.71875,
+      "step": 854
+    },
+    {
+      "epoch": 1.049079754601227,
+      "grad_norm": 12.329495080989613,
+      "learning_rate": 5.256198347107438e-07,
+      "logits/chosen": -0.038818359375,
+      "logits/rejected": -0.185546875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0646,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.84375,
+      "step": 855
+    },
+    {
+      "epoch": 1.0503067484662576,
+      "grad_norm": 13.38637868578377,
+      "learning_rate": 5.252892561983471e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0713,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.09375,
+      "step": 856
+    },
+    {
+      "epoch": 1.0515337423312883,
+      "grad_norm": 9.793442909461401,
+      "learning_rate": 5.249586776859503e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -364.0,
+      "loss": 0.047,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -7.625,
+      "step": 857
+    },
+    {
+      "epoch": 1.052760736196319,
+      "grad_norm": 6.67438475020692,
+      "learning_rate": 5.246280991735537e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.1904296875,
+      "logps/chosen": -318.0,
+      "logps/rejected": -304.0,
+      "loss": 0.0366,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.84375,
+      "step": 858
+    },
+    {
+      "epoch": 1.0539877300613496,
+      "grad_norm": 8.529881979528566,
+      "learning_rate": 5.24297520661157e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0383,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.76953125,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.84375,
+      "step": 859
+    },
+    {
+      "epoch": 1.0552147239263803,
+      "grad_norm": 11.085862425760341,
+      "learning_rate": 5.239669421487603e-07,
+      "logits/chosen": -0.0791015625,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0715,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.46875,
+      "rewards/margins": 7.46875,
+      "rewards/rejected": -7.0,
+      "step": 860
+    },
+    {
+      "epoch": 1.056441717791411,
+      "grad_norm": 9.866985209297734,
+      "learning_rate": 5.236363636363636e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0442,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4296875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -7.59375,
+      "step": 861
+    },
+    {
+      "epoch": 1.0576687116564418,
+      "grad_norm": 9.885391182604536,
+      "learning_rate": 5.233057851239669e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0428,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.36328125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.375,
+      "step": 862
+    },
+    {
+      "epoch": 1.0588957055214725,
+      "grad_norm": 5.796756692523751,
+      "learning_rate": 5.229752066115702e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0241,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.953125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -7.9375,
+      "step": 863
+    },
+    {
+      "epoch": 1.0601226993865032,
+      "grad_norm": 9.27987197358265,
+      "learning_rate": 5.226446280991736e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -368.0,
+      "loss": 0.048,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.5625,
+      "step": 864
+    },
+    {
+      "epoch": 1.0613496932515338,
+      "grad_norm": 8.3395730894386,
+      "learning_rate": 5.223140495867769e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0338,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.859375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.8125,
+      "step": 865
+    },
+    {
+      "epoch": 1.0625766871165645,
+      "grad_norm": 6.886729358178275,
+      "learning_rate": 5.219834710743801e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -348.0,
+      "loss": 0.032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.6875,
+      "step": 866
+    },
+    {
+      "epoch": 1.0638036809815952,
+      "grad_norm": 16.605968137616845,
+      "learning_rate": 5.216528925619834e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0846,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.8125,
+      "step": 867
+    },
+    {
+      "epoch": 1.0650306748466258,
+      "grad_norm": 9.468389714978352,
+      "learning_rate": 5.213223140495867e-07,
+      "logits/chosen": 0.00176239013671875,
+      "logits/rejected": -0.1982421875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -308.0,
+      "loss": 0.0545,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.73046875,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -7.34375,
+      "step": 868
+    },
+    {
+      "epoch": 1.0662576687116565,
+      "grad_norm": 7.444446735409014,
+      "learning_rate": 5.2099173553719e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0385,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.1875,
+      "step": 869
+    },
+    {
+      "epoch": 1.0674846625766872,
+      "grad_norm": 9.293616058876756,
+      "learning_rate": 5.206611570247934e-07,
+      "logits/chosen": -0.06591796875,
+      "logits/rejected": -0.2470703125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0503,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.78125,
+      "step": 870
+    },
+    {
+      "epoch": 1.0687116564417178,
+      "grad_norm": 14.736932461238908,
+      "learning_rate": 5.203305785123967e-07,
+      "logits/chosen": -0.0703125,
+      "logits/rejected": -0.2353515625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0651,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.83984375,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.15625,
+      "step": 871
+    },
+    {
+      "epoch": 1.0699386503067485,
+      "grad_norm": 10.350348873115376,
+      "learning_rate": 5.2e-07,
+      "logits/chosen": -0.087890625,
+      "logits/rejected": -0.2177734375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -318.0,
+      "loss": 0.0463,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.5625,
+      "step": 872
+    },
+    {
+      "epoch": 1.0711656441717792,
+      "grad_norm": 10.092656085910821,
+      "learning_rate": 5.196694214876033e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0513,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.125,
+      "step": 873
+    },
+    {
+      "epoch": 1.0723926380368098,
+      "grad_norm": 18.614352522823648,
+      "learning_rate": 5.193388429752066e-07,
+      "logits/chosen": -0.2001953125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0732,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.71875,
+      "step": 874
+    },
+    {
+      "epoch": 1.0736196319018405,
+      "grad_norm": 7.961276364458803,
+      "learning_rate": 5.190082644628098e-07,
+      "logits/chosen": -0.08203125,
+      "logits/rejected": -0.24609375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0406,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.71875,
+      "step": 875
+    },
+    {
+      "epoch": 1.0748466257668712,
+      "grad_norm": 10.357878513050897,
+      "learning_rate": 5.186776859504132e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0483,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.375,
+      "step": 876
+    },
+    {
+      "epoch": 1.0760736196319018,
+      "grad_norm": 10.744500955796752,
+      "learning_rate": 5.183471074380165e-07,
+      "logits/chosen": -0.0302734375,
+      "logits/rejected": -0.1826171875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.427734375,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.59375,
+      "step": 877
+    },
+    {
+      "epoch": 1.0773006134969325,
+      "grad_norm": 12.618829108696058,
+      "learning_rate": 5.180165289256198e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0761,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.376953125,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -7.75,
+      "step": 878
+    },
+    {
+      "epoch": 1.0785276073619632,
+      "grad_norm": 8.80156369716181,
+      "learning_rate": 5.176859504132231e-07,
+      "logits/chosen": -0.06982421875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0375,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.65625,
+      "step": 879
+    },
+    {
+      "epoch": 1.0797546012269938,
+      "grad_norm": 11.402516366286292,
+      "learning_rate": 5.173553719008264e-07,
+      "logits/chosen": -0.07861328125,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0504,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.062255859375,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.75,
+      "step": 880
+    },
+    {
+      "epoch": 1.0809815950920245,
+      "grad_norm": 11.031397024556938,
+      "learning_rate": 5.170247933884298e-07,
+      "logits/chosen": -0.048828125,
+      "logits/rejected": -0.1884765625,
+      "logps/chosen": -348.0,
+      "logps/rejected": -302.0,
+      "loss": 0.0618,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -6.8125,
+      "step": 881
+    },
+    {
+      "epoch": 1.0822085889570552,
+      "grad_norm": 18.547130207267188,
+      "learning_rate": 5.166942148760331e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0755,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.125,
+      "step": 882
+    },
+    {
+      "epoch": 1.0834355828220859,
+      "grad_norm": 6.6527500793888805,
+      "learning_rate": 5.163636363636364e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0289,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -6.6875,
+      "step": 883
+    },
+    {
+      "epoch": 1.0846625766871165,
+      "grad_norm": 6.861231219217128,
+      "learning_rate": 5.160330578512396e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -426.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0301,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 2.171875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -7.5,
+      "step": 884
+    },
+    {
+      "epoch": 1.0858895705521472,
+      "grad_norm": 9.381597515049744,
+      "learning_rate": 5.157024793388429e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0562,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.78125,
+      "step": 885
+    },
+    {
+      "epoch": 1.0871165644171779,
+      "grad_norm": 8.908837796346583,
+      "learning_rate": 5.153719008264462e-07,
+      "logits/chosen": -0.041748046875,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0518,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 7.75,
+      "rewards/rejected": -6.46875,
+      "step": 886
+    },
+    {
+      "epoch": 1.0883435582822085,
+      "grad_norm": 11.48855243783124,
+      "learning_rate": 5.150413223140496e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0543,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.75,
+      "step": 887
+    },
+    {
+      "epoch": 1.0895705521472392,
+      "grad_norm": 6.793415855369003,
+      "learning_rate": 5.147107438016529e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0336,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.34375,
+      "step": 888
+    },
+    {
+      "epoch": 1.0907975460122699,
+      "grad_norm": 15.354710995482474,
+      "learning_rate": 5.143801652892562e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -450.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0741,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.0625,
+      "step": 889
+    },
+    {
+      "epoch": 1.0920245398773005,
+      "grad_norm": 9.766896381796215,
+      "learning_rate": 5.140495867768595e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0492,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -7.0,
+      "step": 890
+    },
+    {
+      "epoch": 1.0932515337423312,
+      "grad_norm": 9.533700281126114,
+      "learning_rate": 5.137190082644628e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0521,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.8125,
+      "step": 891
+    },
+    {
+      "epoch": 1.0944785276073619,
+      "grad_norm": 8.94436680718413,
+      "learning_rate": 5.133884297520661e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -340.0,
+      "loss": 0.049,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.875,
+      "step": 892
+    },
+    {
+      "epoch": 1.0957055214723925,
+      "grad_norm": 13.394144993964538,
+      "learning_rate": 5.130578512396694e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.2099609375,
+      "logps/chosen": -316.0,
+      "logps/rejected": -312.0,
+      "loss": 0.0707,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 0.8125,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.1875,
+      "step": 893
+    },
+    {
+      "epoch": 1.0969325153374232,
+      "grad_norm": 17.301836509277987,
+      "learning_rate": 5.127272727272727e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0535,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.7265625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.125,
+      "step": 894
+    },
+    {
+      "epoch": 1.098159509202454,
+      "grad_norm": 13.580959238382313,
+      "learning_rate": 5.12396694214876e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0519,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -7.03125,
+      "step": 895
+    },
+    {
+      "epoch": 1.0993865030674848,
+      "grad_norm": 10.698325211425429,
+      "learning_rate": 5.120661157024793e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0481,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.46875,
+      "step": 896
+    },
+    {
+      "epoch": 1.1006134969325154,
+      "grad_norm": 11.842993344318062,
+      "learning_rate": 5.117355371900826e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -440.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0496,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.84375,
+      "step": 897
+    },
+    {
+      "epoch": 1.101840490797546,
+      "grad_norm": 7.993601971885361,
+      "learning_rate": 5.114049586776859e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0393,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.0625,
+      "step": 898
+    },
+    {
+      "epoch": 1.1030674846625768,
+      "grad_norm": 7.485760340921191,
+      "learning_rate": 5.110743801652893e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0429,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.287109375,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -8.0625,
+      "step": 899
+    },
+    {
+      "epoch": 1.1042944785276074,
+      "grad_norm": 10.22220098269188,
+      "learning_rate": 5.107438016528926e-07,
+      "logits/chosen": -0.10205078125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -380.0,
+      "loss": 0.046,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.80859375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.875,
+      "step": 900
+    },
+    {
+      "epoch": 1.105521472392638,
+      "grad_norm": 11.370160710024297,
+      "learning_rate": 5.104132231404959e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -490.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0534,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.828125,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -8.6875,
+      "step": 901
+    },
+    {
+      "epoch": 1.1067484662576688,
+      "grad_norm": 12.269173865343777,
+      "learning_rate": 5.100826446280991e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0518,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.91015625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.5625,
+      "step": 902
+    },
+    {
+      "epoch": 1.1079754601226994,
+      "grad_norm": 10.521897818355166,
+      "learning_rate": 5.097520661157024e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0463,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.1875,
+      "step": 903
+    },
+    {
+      "epoch": 1.10920245398773,
+      "grad_norm": 6.33098217292226,
+      "learning_rate": 5.094214876033057e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0288,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.5,
+      "step": 904
+    },
+    {
+      "epoch": 1.1104294478527608,
+      "grad_norm": 10.008456465270969,
+      "learning_rate": 5.090909090909091e-07,
+      "logits/chosen": -0.330078125,
+      "logits/rejected": -0.54296875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0393,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -8.4375,
+      "step": 905
+    },
+    {
+      "epoch": 1.1116564417177914,
+      "grad_norm": 9.946359264363899,
+      "learning_rate": 5.087603305785124e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -336.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0528,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.126953125,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -7.90625,
+      "step": 906
+    },
+    {
+      "epoch": 1.112883435582822,
+      "grad_norm": 13.90671388757752,
+      "learning_rate": 5.084297520661157e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0595,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.625,
+      "step": 907
+    },
+    {
+      "epoch": 1.1141104294478528,
+      "grad_norm": 17.130329509275274,
+      "learning_rate": 5.08099173553719e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0457,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.78125,
+      "step": 908
+    },
+    {
+      "epoch": 1.1153374233128834,
+      "grad_norm": 13.27945349361827,
+      "learning_rate": 5.077685950413223e-07,
+      "logits/chosen": -0.11572265625,
+      "logits/rejected": -0.228515625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0579,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.375,
+      "step": 909
+    },
+    {
+      "epoch": 1.116564417177914,
+      "grad_norm": 9.737802290234768,
+      "learning_rate": 5.074380165289256e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0329,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.375,
+      "step": 910
+    },
+    {
+      "epoch": 1.1177914110429448,
+      "grad_norm": 6.685651209591217,
+      "learning_rate": 5.071074380165289e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -466.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0268,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -9.0,
+      "step": 911
+    },
+    {
+      "epoch": 1.1190184049079754,
+      "grad_norm": 13.867209331156516,
+      "learning_rate": 5.067768595041322e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.19140625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0517,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.8359375,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.8125,
+      "step": 912
+    },
+    {
+      "epoch": 1.120245398773006,
+      "grad_norm": 11.283456894779777,
+      "learning_rate": 5.064462809917355e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -322.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0591,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.625,
+      "step": 913
+    },
+    {
+      "epoch": 1.1214723926380368,
+      "grad_norm": 13.891782784573822,
+      "learning_rate": 5.061157024793388e-07,
+      "logits/chosen": -0.09423828125,
+      "logits/rejected": -0.26953125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0699,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.625,
+      "step": 914
+    },
+    {
+      "epoch": 1.1226993865030674,
+      "grad_norm": 15.737683653007828,
+      "learning_rate": 5.057851239669421e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0565,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.0,
+      "step": 915
+    },
+    {
+      "epoch": 1.123926380368098,
+      "grad_norm": 9.209835377946382,
+      "learning_rate": 5.054545454545454e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0438,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.203125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.90625,
+      "step": 916
+    },
+    {
+      "epoch": 1.1251533742331288,
+      "grad_norm": 14.318300360080142,
+      "learning_rate": 5.051239669421488e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -404.0,
+      "loss": 0.072,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.875,
+      "step": 917
+    },
+    {
+      "epoch": 1.1263803680981594,
+      "grad_norm": 16.551514820318832,
+      "learning_rate": 5.047933884297521e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0695,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.28125,
+      "step": 918
+    },
+    {
+      "epoch": 1.1276073619631901,
+      "grad_norm": 11.082245597349267,
+      "learning_rate": 5.044628099173554e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0584,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.90625,
+      "step": 919
+    },
+    {
+      "epoch": 1.1288343558282208,
+      "grad_norm": 12.619871004485745,
+      "learning_rate": 5.041322314049586e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0433,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -7.71875,
+      "step": 920
+    },
+    {
+      "epoch": 1.1300613496932514,
+      "grad_norm": 10.77826696391865,
+      "learning_rate": 5.038016528925619e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0345,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.6484375,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.6875,
+      "step": 921
+    },
+    {
+      "epoch": 1.1312883435582821,
+      "grad_norm": 9.71506841489295,
+      "learning_rate": 5.034710743801652e-07,
+      "logits/chosen": -0.12158203125,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0473,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.9140625,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -6.875,
+      "step": 922
+    },
+    {
+      "epoch": 1.132515337423313,
+      "grad_norm": 8.986837909388852,
+      "learning_rate": 5.031404958677686e-07,
+      "logits/chosen": -0.1064453125,
+      "logits/rejected": -0.2041015625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0394,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 8.3125,
+      "rewards/rejected": -6.71875,
+      "step": 923
+    },
+    {
+      "epoch": 1.1337423312883437,
+      "grad_norm": 11.240488448077638,
+      "learning_rate": 5.028099173553719e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -340.0,
+      "loss": 0.048,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -7.53125,
+      "step": 924
+    },
+    {
+      "epoch": 1.1349693251533743,
+      "grad_norm": 10.37676153526059,
+      "learning_rate": 5.024793388429752e-07,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.5,
+      "logps/chosen": -372.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0385,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.66796875,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -7.90625,
+      "step": 925
+    },
+    {
+      "epoch": 1.136196319018405,
+      "grad_norm": 14.556750345009446,
+      "learning_rate": 5.021487603305785e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0586,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.859375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -7.5625,
+      "step": 926
+    },
+    {
+      "epoch": 1.1374233128834357,
+      "grad_norm": 8.6263295939874,
+      "learning_rate": 5.018181818181818e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0309,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.625,
+      "step": 927
+    },
+    {
+      "epoch": 1.1386503067484663,
+      "grad_norm": 7.620881147688411,
+      "learning_rate": 5.014876033057852e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0414,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.53125,
+      "step": 928
+    },
+    {
+      "epoch": 1.139877300613497,
+      "grad_norm": 11.937716031232593,
+      "learning_rate": 5.011570247933884e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -426.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0443,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.4375,
+      "step": 929
+    },
+    {
+      "epoch": 1.1411042944785277,
+      "grad_norm": 25.46977596486142,
+      "learning_rate": 5.008264462809917e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0556,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.25,
+      "step": 930
+    },
+    {
+      "epoch": 1.1423312883435583,
+      "grad_norm": 12.357953826396704,
+      "learning_rate": 5.00495867768595e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0626,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.7265625,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -7.1875,
+      "step": 931
+    },
+    {
+      "epoch": 1.143558282208589,
+      "grad_norm": 10.124449885714515,
+      "learning_rate": 5.001652892561983e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0487,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.87109375,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.75,
+      "step": 932
+    },
+    {
+      "epoch": 1.1447852760736197,
+      "grad_norm": 10.186523170112832,
+      "learning_rate": 4.998347107438016e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0407,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.53125,
+      "step": 933
+    },
+    {
+      "epoch": 1.1460122699386504,
+      "grad_norm": 14.784919527567176,
+      "learning_rate": 4.99504132231405e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -344.0,
+      "loss": 0.055,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.09375,
+      "step": 934
+    },
+    {
+      "epoch": 1.147239263803681,
+      "grad_norm": 8.084085375609561,
+      "learning_rate": 4.991735537190083e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0339,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.5625,
+      "step": 935
+    },
+    {
+      "epoch": 1.1484662576687117,
+      "grad_norm": 10.090762024483787,
+      "learning_rate": 4.988429752066116e-07,
+      "logits/chosen": -0.06787109375,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0428,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.66796875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.4375,
+      "step": 936
+    },
+    {
+      "epoch": 1.1496932515337424,
+      "grad_norm": 11.391883354259097,
+      "learning_rate": 4.985123966942148e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0446,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.4375,
+      "step": 937
+    },
+    {
+      "epoch": 1.150920245398773,
+      "grad_norm": 17.337934806380964,
+      "learning_rate": 4.981818181818181e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.2001953125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0702,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.73828125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.25,
+      "step": 938
+    },
+    {
+      "epoch": 1.1521472392638037,
+      "grad_norm": 6.607905978535942,
+      "learning_rate": 4.978512396694214e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -444.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0298,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -8.8125,
+      "step": 939
+    },
+    {
+      "epoch": 1.1533742331288344,
+      "grad_norm": 7.07231413487028,
+      "learning_rate": 4.975206611570248e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0283,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.0,
+      "step": 940
+    },
+    {
+      "epoch": 1.154601226993865,
+      "grad_norm": 14.928874081631259,
+      "learning_rate": 4.971900826446281e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0622,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.71875,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -7.8125,
+      "step": 941
+    },
+    {
+      "epoch": 1.1558282208588957,
+      "grad_norm": 17.516711144244283,
+      "learning_rate": 4.968595041322314e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0501,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -9.3125,
+      "step": 942
+    },
+    {
+      "epoch": 1.1570552147239264,
+      "grad_norm": 8.961156359700125,
+      "learning_rate": 4.965289256198347e-07,
+      "logits/chosen": -0.10400390625,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0371,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.62109375,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.0,
+      "step": 943
+    },
+    {
+      "epoch": 1.158282208588957,
+      "grad_norm": 8.12773430831165,
+      "learning_rate": 4.96198347107438e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0336,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.75,
+      "step": 944
+    },
+    {
+      "epoch": 1.1595092024539877,
+      "grad_norm": 11.434978350030235,
+      "learning_rate": 4.958677685950413e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0455,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.0,
+      "step": 945
+    },
+    {
+      "epoch": 1.1607361963190184,
+      "grad_norm": 10.093045663484858,
+      "learning_rate": 4.955371900826446e-07,
+      "logits/chosen": -0.1240234375,
+      "logits/rejected": -0.205078125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0429,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.53125,
+      "step": 946
+    },
+    {
+      "epoch": 1.161963190184049,
+      "grad_norm": 11.765426924290539,
+      "learning_rate": 4.952066115702479e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -342.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0689,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -7.5,
+      "step": 947
+    },
+    {
+      "epoch": 1.1631901840490797,
+      "grad_norm": 12.622464974038337,
+      "learning_rate": 4.948760330578512e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0432,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -7.46875,
+      "step": 948
+    },
+    {
+      "epoch": 1.1644171779141104,
+      "grad_norm": 11.580415695153775,
+      "learning_rate": 4.945454545454545e-07,
+      "logits/chosen": -0.119140625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0559,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -7.6875,
+      "step": 949
+    },
+    {
+      "epoch": 1.165644171779141,
+      "grad_norm": 9.115099300613052,
+      "learning_rate": 4.942148760330578e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -334.0,
+      "logps/rejected": -310.0,
+      "loss": 0.0411,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.7890625,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.65625,
+      "step": 950
+    },
+    {
+      "epoch": 1.1668711656441717,
+      "grad_norm": 11.047710529408716,
+      "learning_rate": 4.938842975206611e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0529,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.73046875,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -7.40625,
+      "step": 951
+    },
+    {
+      "epoch": 1.1680981595092024,
+      "grad_norm": 13.283612267053757,
+      "learning_rate": 4.935537190082645e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0633,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.84375,
+      "step": 952
+    },
+    {
+      "epoch": 1.169325153374233,
+      "grad_norm": 12.688088508685484,
+      "learning_rate": 4.932231404958678e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0515,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -7.0625,
+      "step": 953
+    },
+    {
+      "epoch": 1.1705521472392637,
+      "grad_norm": 13.72625399518082,
+      "learning_rate": 4.928925619834711e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -316.0,
+      "logps/rejected": -294.0,
+      "loss": 0.0694,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.17578125,
+      "rewards/margins": 7.84375,
+      "rewards/rejected": -7.6875,
+      "step": 954
+    },
+    {
+      "epoch": 1.1717791411042944,
+      "grad_norm": 15.264557868135348,
+      "learning_rate": 4.925619834710743e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.2080078125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0505,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4296875,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.25,
+      "step": 955
+    },
+    {
+      "epoch": 1.173006134969325,
+      "grad_norm": 8.222673231264135,
+      "learning_rate": 4.922314049586776e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -310.0,
+      "logps/rejected": -320.0,
+      "loss": 0.0373,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 7.875,
+      "rewards/rejected": -7.125,
+      "step": 956
+    },
+    {
+      "epoch": 1.1742331288343557,
+      "grad_norm": 8.70101877312667,
+      "learning_rate": 4.919008264462809e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0379,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.34375,
+      "step": 957
+    },
+    {
+      "epoch": 1.1754601226993864,
+      "grad_norm": 8.625233525867843,
+      "learning_rate": 4.915702479338843e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0329,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.0625,
+      "step": 958
+    },
+    {
+      "epoch": 1.1766871165644173,
+      "grad_norm": 8.831414998360419,
+      "learning_rate": 4.912396694214876e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0393,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.71875,
+      "step": 959
+    },
+    {
+      "epoch": 1.177914110429448,
+      "grad_norm": 7.156930699104763,
+      "learning_rate": 4.909090909090909e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0348,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.265625,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -8.25,
+      "step": 960
+    },
+    {
+      "epoch": 1.1791411042944786,
+      "grad_norm": 7.316598874843914,
+      "learning_rate": 4.905785123966942e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -356.0,
+      "loss": 0.032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.125,
+      "step": 961
+    },
+    {
+      "epoch": 1.1803680981595093,
+      "grad_norm": 15.127473084571868,
+      "learning_rate": 4.902479338842975e-07,
+      "logits/chosen": -0.1220703125,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0505,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.96875,
+      "step": 962
+    },
+    {
+      "epoch": 1.18159509202454,
+      "grad_norm": 7.350557070973821,
+      "learning_rate": 4.899173553719008e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0355,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -7.90625,
+      "step": 963
+    },
+    {
+      "epoch": 1.1828220858895706,
+      "grad_norm": 7.9216285170357805,
+      "learning_rate": 4.895867768595041e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0269,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.25,
+      "step": 964
+    },
+    {
+      "epoch": 1.1840490797546013,
+      "grad_norm": 17.652292078469603,
+      "learning_rate": 4.892561983471074e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0483,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.55078125,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.5,
+      "step": 965
+    },
+    {
+      "epoch": 1.185276073619632,
+      "grad_norm": 9.347424143338342,
+      "learning_rate": 4.889256198347107e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0449,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.83203125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.1875,
+      "step": 966
+    },
+    {
+      "epoch": 1.1865030674846626,
+      "grad_norm": 8.997313851138234,
+      "learning_rate": 4.88595041322314e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0435,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.71875,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.0,
+      "step": 967
+    },
+    {
+      "epoch": 1.1877300613496933,
+      "grad_norm": 8.350460284446099,
+      "learning_rate": 4.882644628099173e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0278,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -8.9375,
+      "step": 968
+    },
+    {
+      "epoch": 1.188957055214724,
+      "grad_norm": 17.12432481280222,
+      "learning_rate": 4.879338842975206e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -396.0,
+      "loss": 0.063,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.453125,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -8.6875,
+      "step": 969
+    },
+    {
+      "epoch": 1.1901840490797546,
+      "grad_norm": 11.73493140010646,
+      "learning_rate": 4.87603305785124e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -484.0,
+      "loss": 0.0539,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 22.0,
+      "rewards/rejected": -21.25,
+      "step": 970
+    },
+    {
+      "epoch": 1.1914110429447853,
+      "grad_norm": 11.251800364854649,
+      "learning_rate": 4.872727272727273e-07,
+      "logits/chosen": -0.1357421875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0577,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.50390625,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.4375,
+      "step": 971
+    },
+    {
+      "epoch": 1.192638036809816,
+      "grad_norm": 12.538848815995712,
+      "learning_rate": 4.869421487603306e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -338.0,
+      "logps/rejected": -292.0,
+      "loss": 0.0648,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1181640625,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -7.875,
+      "step": 972
+    },
+    {
+      "epoch": 1.1938650306748466,
+      "grad_norm": 8.47785010044017,
+      "learning_rate": 4.866115702479338e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0277,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.5,
+      "step": 973
+    },
+    {
+      "epoch": 1.1950920245398773,
+      "grad_norm": 9.052453334123884,
+      "learning_rate": 4.862809917355371e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0343,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -8.875,
+      "step": 974
+    },
+    {
+      "epoch": 1.196319018404908,
+      "grad_norm": 12.01099472634397,
+      "learning_rate": 4.859504132231405e-07,
+      "logits/chosen": -0.095703125,
+      "logits/rejected": -0.22265625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -314.0,
+      "loss": 0.041,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.78125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -7.96875,
+      "step": 975
+    },
+    {
+      "epoch": 1.1975460122699386,
+      "grad_norm": 11.988662613154055,
+      "learning_rate": 4.856198347107438e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0538,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.375,
+      "step": 976
+    },
+    {
+      "epoch": 1.1987730061349693,
+      "grad_norm": 11.13765858997807,
+      "learning_rate": 4.852892561983471e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0288,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -8.8125,
+      "step": 977
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 13.793986032637335,
+      "learning_rate": 4.849586776859504e-07,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0739,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.3125,
+      "step": 978
+    },
+    {
+      "epoch": 1.2012269938650306,
+      "grad_norm": 8.480774562673568,
+      "learning_rate": 4.846280991735537e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0307,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.4375,
+      "step": 979
+    },
+    {
+      "epoch": 1.2024539877300613,
+      "grad_norm": 4.365430571481909,
+      "learning_rate": 4.84297520661157e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0198,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.0625,
+      "step": 980
+    },
+    {
+      "epoch": 1.203680981595092,
+      "grad_norm": 13.729307845818116,
+      "learning_rate": 4.839669421487604e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0469,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.5625,
+      "step": 981
+    },
+    {
+      "epoch": 1.2049079754601226,
+      "grad_norm": 10.207687106718279,
+      "learning_rate": 4.836363636363636e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0494,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.875,
+      "step": 982
+    },
+    {
+      "epoch": 1.2061349693251533,
+      "grad_norm": 17.058047206794242,
+      "learning_rate": 4.833057851239669e-07,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0847,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -6.8125,
+      "step": 983
+    },
+    {
+      "epoch": 1.207361963190184,
+      "grad_norm": 12.189727136679215,
+      "learning_rate": 4.829752066115702e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0584,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.75,
+      "step": 984
+    },
+    {
+      "epoch": 1.2085889570552146,
+      "grad_norm": 12.801014204687064,
+      "learning_rate": 4.826446280991735e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -344.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0543,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.46875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.375,
+      "step": 985
+    },
+    {
+      "epoch": 1.2098159509202455,
+      "grad_norm": 9.609094722284011,
+      "learning_rate": 4.823140495867768e-07,
+      "logits/chosen": -0.050537109375,
+      "logits/rejected": -0.1962890625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0361,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -6.875,
+      "step": 986
+    },
+    {
+      "epoch": 1.2110429447852762,
+      "grad_norm": 13.218029076540835,
+      "learning_rate": 4.819834710743802e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -452.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0483,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -7.78125,
+      "step": 987
+    },
+    {
+      "epoch": 1.2122699386503069,
+      "grad_norm": 8.141541687744194,
+      "learning_rate": 4.816528925619835e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0309,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.9921875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.9375,
+      "step": 988
+    },
+    {
+      "epoch": 1.2134969325153375,
+      "grad_norm": 12.298643628948078,
+      "learning_rate": 4.813223140495868e-07,
+      "logits/chosen": -0.107421875,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -318.0,
+      "loss": 0.0631,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.67578125,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -7.40625,
+      "step": 989
+    },
+    {
+      "epoch": 1.2147239263803682,
+      "grad_norm": 5.79863996841087,
+      "learning_rate": 4.809917355371901e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0325,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.625,
+      "step": 990
+    },
+    {
+      "epoch": 1.2159509202453989,
+      "grad_norm": 6.760236470696033,
+      "learning_rate": 4.806611570247933e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0284,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.0625,
+      "step": 991
+    },
+    {
+      "epoch": 1.2171779141104295,
+      "grad_norm": 10.672410560638603,
+      "learning_rate": 4.803305785123966e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0472,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.5625,
+      "step": 992
+    },
+    {
+      "epoch": 1.2184049079754602,
+      "grad_norm": 9.716010918159654,
+      "learning_rate": 4.8e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0385,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -7.96875,
+      "step": 993
+    },
+    {
+      "epoch": 1.2196319018404909,
+      "grad_norm": 10.42174132730132,
+      "learning_rate": 4.796694214876033e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0418,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.625,
+      "step": 994
+    },
+    {
+      "epoch": 1.2208588957055215,
+      "grad_norm": 11.123972657771333,
+      "learning_rate": 4.793388429752066e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0373,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.462890625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.75,
+      "step": 995
+    },
+    {
+      "epoch": 1.2220858895705522,
+      "grad_norm": 25.75709421241266,
+      "learning_rate": 4.790082644628099e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0693,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.41015625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.4375,
+      "step": 996
+    },
+    {
+      "epoch": 1.2233128834355829,
+      "grad_norm": 9.409931474479396,
+      "learning_rate": 4.786776859504132e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0363,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.0,
+      "step": 997
+    },
+    {
+      "epoch": 1.2245398773006135,
+      "grad_norm": 9.490599131683323,
+      "learning_rate": 4.783471074380165e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -442.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0311,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.125,
+      "step": 998
+    },
+    {
+      "epoch": 1.2257668711656442,
+      "grad_norm": 8.127599963420865,
+      "learning_rate": 4.780165289256199e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0319,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.25,
+      "step": 999
+    },
+    {
+      "epoch": 1.2269938650306749,
+      "grad_norm": 12.601959839650176,
+      "learning_rate": 4.776859504132231e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0659,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -7.8125,
+      "step": 1000
+    },
+    {
+      "epoch": 1.2282208588957055,
+      "grad_norm": 10.056271597433025,
+      "learning_rate": 4.773553719008264e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0356,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.3125,
+      "step": 1001
+    },
+    {
+      "epoch": 1.2294478527607362,
+      "grad_norm": 17.964616987432837,
+      "learning_rate": 4.770247933884297e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0573,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -7.96875,
+      "step": 1002
+    },
+    {
+      "epoch": 1.2306748466257669,
+      "grad_norm": 10.775894884702035,
+      "learning_rate": 4.7669421487603303e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0485,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.92578125,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.25,
+      "step": 1003
+    },
+    {
+      "epoch": 1.2319018404907975,
+      "grad_norm": 8.116589015572577,
+      "learning_rate": 4.7636363636363635e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -308.0,
+      "loss": 0.0375,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.375,
+      "step": 1004
+    },
+    {
+      "epoch": 1.2331288343558282,
+      "grad_norm": 10.13931784187266,
+      "learning_rate": 4.7603305785123966e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0502,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.25,
+      "step": 1005
+    },
+    {
+      "epoch": 1.2343558282208589,
+      "grad_norm": 8.80736282586411,
+      "learning_rate": 4.75702479338843e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0403,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.375,
+      "step": 1006
+    },
+    {
+      "epoch": 1.2355828220858895,
+      "grad_norm": 9.046118916104152,
+      "learning_rate": 4.753719008264463e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0411,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.3125,
+      "step": 1007
+    },
+    {
+      "epoch": 1.2368098159509202,
+      "grad_norm": 12.220589771919323,
+      "learning_rate": 4.750413223140496e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0435,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.59375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.0625,
+      "step": 1008
+    },
+    {
+      "epoch": 1.2380368098159509,
+      "grad_norm": 14.688018091941988,
+      "learning_rate": 4.7471074380165283e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0554,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3828125,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -7.8125,
+      "step": 1009
+    },
+    {
+      "epoch": 1.2392638036809815,
+      "grad_norm": 11.98135604057609,
+      "learning_rate": 4.7438016528925615e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0605,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.032470703125,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.96875,
+      "step": 1010
+    },
+    {
+      "epoch": 1.2404907975460122,
+      "grad_norm": 11.577259622901868,
+      "learning_rate": 4.7404958677685946e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0444,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.64453125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.3125,
+      "step": 1011
+    },
+    {
+      "epoch": 1.2417177914110429,
+      "grad_norm": 16.548853860409746,
+      "learning_rate": 4.737190082644628e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0828,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.8359375,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.0,
+      "step": 1012
+    },
+    {
+      "epoch": 1.2429447852760735,
+      "grad_norm": 8.191313021502108,
+      "learning_rate": 4.733884297520661e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.2265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0457,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.37890625,
+      "rewards/margins": 7.875,
+      "rewards/rejected": -7.5,
+      "step": 1013
+    },
+    {
+      "epoch": 1.2441717791411042,
+      "grad_norm": 10.841806442484222,
+      "learning_rate": 4.730578512396694e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.388671875,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -8.0625,
+      "step": 1014
+    },
+    {
+      "epoch": 1.2453987730061349,
+      "grad_norm": 12.095214544652428,
+      "learning_rate": 4.7272727272727273e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.0625,
+      "step": 1015
+    },
+    {
+      "epoch": 1.2466257668711656,
+      "grad_norm": 20.048615612397384,
+      "learning_rate": 4.7239669421487605e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0901,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.44921875,
+      "rewards/margins": 7.875,
+      "rewards/rejected": -7.4375,
+      "step": 1016
+    },
+    {
+      "epoch": 1.2478527607361962,
+      "grad_norm": 4.909314153216152,
+      "learning_rate": 4.7206611570247937e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0205,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.125,
+      "step": 1017
+    },
+    {
+      "epoch": 1.2490797546012269,
+      "grad_norm": 15.913032240180387,
+      "learning_rate": 4.717355371900826e-07,
+      "logits/chosen": -0.248046875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0537,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.46484375,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.1875,
+      "step": 1018
+    },
+    {
+      "epoch": 1.2503067484662576,
+      "grad_norm": 7.998330769581237,
+      "learning_rate": 4.714049586776859e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0435,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7421875,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.8125,
+      "step": 1019
+    },
+    {
+      "epoch": 1.2515337423312882,
+      "grad_norm": 8.154520742857182,
+      "learning_rate": 4.710743801652892e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0424,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.72265625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.375,
+      "step": 1020
+    },
+    {
+      "epoch": 1.252760736196319,
+      "grad_norm": 10.190607926042517,
+      "learning_rate": 4.7074380165289254e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0457,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.216796875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.8125,
+      "step": 1021
+    },
+    {
+      "epoch": 1.2539877300613496,
+      "grad_norm": 9.551253572647711,
+      "learning_rate": 4.7041322314049585e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0398,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.76171875,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.1875,
+      "step": 1022
+    },
+    {
+      "epoch": 1.2552147239263804,
+      "grad_norm": 9.671012064033086,
+      "learning_rate": 4.7008264462809917e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0413,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -8.875,
+      "step": 1023
+    },
+    {
+      "epoch": 1.2564417177914111,
+      "grad_norm": 12.47562958036183,
+      "learning_rate": 4.697520661157025e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0505,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.5,
+      "step": 1024
+    },
+    {
+      "epoch": 1.2576687116564418,
+      "grad_norm": 12.875244309972366,
+      "learning_rate": 4.694214876033058e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0622,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -8.6875,
+      "step": 1025
+    },
+    {
+      "epoch": 1.2588957055214725,
+      "grad_norm": 9.287273683982628,
+      "learning_rate": 4.69090909090909e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -362.0,
+      "loss": 0.047,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.361328125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.125,
+      "step": 1026
+    },
+    {
+      "epoch": 1.2601226993865031,
+      "grad_norm": 8.982752724660957,
+      "learning_rate": 4.6876033057851234e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -446.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0343,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.78515625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.375,
+      "step": 1027
+    },
+    {
+      "epoch": 1.2613496932515338,
+      "grad_norm": 16.331351941268387,
+      "learning_rate": 4.6842975206611565e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0478,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -9.375,
+      "step": 1028
+    },
+    {
+      "epoch": 1.2625766871165645,
+      "grad_norm": 14.883928893155497,
+      "learning_rate": 4.6809917355371897e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0462,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.5,
+      "step": 1029
+    },
+    {
+      "epoch": 1.2638036809815951,
+      "grad_norm": 8.911098912817012,
+      "learning_rate": 4.677685950413223e-07,
+      "logits/chosen": -0.38671875,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0305,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.4140625,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -9.25,
+      "step": 1030
+    },
+    {
+      "epoch": 1.2650306748466258,
+      "grad_norm": 11.44985669629329,
+      "learning_rate": 4.674380165289256e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0695,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.8359375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.5,
+      "step": 1031
+    },
+    {
+      "epoch": 1.2662576687116565,
+      "grad_norm": 6.783926726533355,
+      "learning_rate": 4.671074380165289e-07,
+      "logits/chosen": -0.07470703125,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0295,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.4375,
+      "step": 1032
+    },
+    {
+      "epoch": 1.2674846625766871,
+      "grad_norm": 13.190345224116596,
+      "learning_rate": 4.6677685950413224e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -430.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0611,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.78515625,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.1875,
+      "step": 1033
+    },
+    {
+      "epoch": 1.2687116564417178,
+      "grad_norm": 6.185510880413833,
+      "learning_rate": 4.6644628099173556e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -460.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0256,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -8.8125,
+      "step": 1034
+    },
+    {
+      "epoch": 1.2699386503067485,
+      "grad_norm": 7.619537783501127,
+      "learning_rate": 4.6611570247933877e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0297,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.625,
+      "step": 1035
+    },
+    {
+      "epoch": 1.2711656441717791,
+      "grad_norm": 13.624124945358417,
+      "learning_rate": 4.657851239669421e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0562,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.65625,
+      "step": 1036
+    },
+    {
+      "epoch": 1.2723926380368098,
+      "grad_norm": 7.586885211554758,
+      "learning_rate": 4.654545454545454e-07,
+      "logits/chosen": -0.35546875,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0282,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.96875,
+      "step": 1037
+    },
+    {
+      "epoch": 1.2736196319018405,
+      "grad_norm": 9.803215310550355,
+      "learning_rate": 4.651239669421487e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0361,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.5,
+      "step": 1038
+    },
+    {
+      "epoch": 1.2748466257668711,
+      "grad_norm": 14.589804231321581,
+      "learning_rate": 4.6479338842975204e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0627,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.8359375,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.59375,
+      "step": 1039
+    },
+    {
+      "epoch": 1.2760736196319018,
+      "grad_norm": 13.800642679130199,
+      "learning_rate": 4.6446280991735536e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0433,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.46875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.375,
+      "step": 1040
+    },
+    {
+      "epoch": 1.2773006134969325,
+      "grad_norm": 6.6360914407229785,
+      "learning_rate": 4.641322314049587e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0192,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.98828125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.125,
+      "step": 1041
+    },
+    {
+      "epoch": 1.2785276073619631,
+      "grad_norm": 10.335577733951094,
+      "learning_rate": 4.63801652892562e-07,
+      "logits/chosen": -0.369140625,
+      "logits/rejected": -0.57421875,
+      "logps/chosen": -444.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0391,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.6875,
+      "step": 1042
+    },
+    {
+      "epoch": 1.2797546012269938,
+      "grad_norm": 20.153647410714772,
+      "learning_rate": 4.634710743801653e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0901,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 0.06396484375,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -7.90625,
+      "step": 1043
+    },
+    {
+      "epoch": 1.2809815950920245,
+      "grad_norm": 6.924901551687326,
+      "learning_rate": 4.631404958677685e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -374.0,
+      "loss": 0.022,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -8.3125,
+      "step": 1044
+    },
+    {
+      "epoch": 1.2822085889570551,
+      "grad_norm": 12.95174867862595,
+      "learning_rate": 4.6280991735537184e-07,
+      "logits/chosen": -0.31640625,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0403,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.125,
+      "step": 1045
+    },
+    {
+      "epoch": 1.2834355828220858,
+      "grad_norm": 14.546745301507391,
+      "learning_rate": 4.6247933884297516e-07,
+      "logits/chosen": -0.099609375,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -374.0,
+      "loss": 0.063,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.83984375,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.53125,
+      "step": 1046
+    },
+    {
+      "epoch": 1.2846625766871167,
+      "grad_norm": 10.883403162683157,
+      "learning_rate": 4.621487603305785e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -376.0,
+      "loss": 0.035,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -8.625,
+      "step": 1047
+    },
+    {
+      "epoch": 1.2858895705521474,
+      "grad_norm": 8.026266419876157,
+      "learning_rate": 4.618181818181818e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -332.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0379,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.375,
+      "step": 1048
+    },
+    {
+      "epoch": 1.287116564417178,
+      "grad_norm": 9.940422173754929,
+      "learning_rate": 4.614876033057851e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0394,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.3125,
+      "step": 1049
+    },
+    {
+      "epoch": 1.2883435582822087,
+      "grad_norm": 13.289809124549075,
+      "learning_rate": 4.6115702479338843e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0543,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.25,
+      "step": 1050
+    },
+    {
+      "epoch": 1.2895705521472394,
+      "grad_norm": 8.246230191346283,
+      "learning_rate": 4.6082644628099175e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0325,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -8.875,
+      "step": 1051
+    },
+    {
+      "epoch": 1.29079754601227,
+      "grad_norm": 14.40820979773154,
+      "learning_rate": 4.6049586776859507e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0636,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -7.9375,
+      "step": 1052
+    },
+    {
+      "epoch": 1.2920245398773007,
+      "grad_norm": 11.868252173578338,
+      "learning_rate": 4.601652892561983e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0532,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.25,
+      "step": 1053
+    },
+    {
+      "epoch": 1.2932515337423314,
+      "grad_norm": 6.870935187145037,
+      "learning_rate": 4.598347107438016e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0361,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.4375,
+      "step": 1054
+    },
+    {
+      "epoch": 1.294478527607362,
+      "grad_norm": 10.48304399523796,
+      "learning_rate": 4.595041322314049e-07,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0388,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.8203125,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.875,
+      "step": 1055
+    },
+    {
+      "epoch": 1.2957055214723927,
+      "grad_norm": 4.362691885640252,
+      "learning_rate": 4.5917355371900823e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -322.0,
+      "loss": 0.0164,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.7734375,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.5,
+      "step": 1056
+    },
+    {
+      "epoch": 1.2969325153374234,
+      "grad_norm": 15.040919629551425,
+      "learning_rate": 4.5884297520661155e-07,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0419,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.375,
+      "step": 1057
+    },
+    {
+      "epoch": 1.298159509202454,
+      "grad_norm": 16.109841585162464,
+      "learning_rate": 4.5851239669421487e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0663,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -8.6875,
+      "step": 1058
+    },
+    {
+      "epoch": 1.2993865030674847,
+      "grad_norm": 7.994341213442368,
+      "learning_rate": 4.581818181818182e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0447,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.80859375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.875,
+      "step": 1059
+    },
+    {
+      "epoch": 1.3006134969325154,
+      "grad_norm": 14.221171845997816,
+      "learning_rate": 4.578512396694215e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -332.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0483,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.625,
+      "step": 1060
+    },
+    {
+      "epoch": 1.301840490797546,
+      "grad_norm": 11.819665312091933,
+      "learning_rate": 4.575206611570248e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0422,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.25,
+      "step": 1061
+    },
+    {
+      "epoch": 1.3030674846625767,
+      "grad_norm": 14.568973750937094,
+      "learning_rate": 4.5719008264462803e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -392.0,
+      "loss": 0.066,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.1875,
+      "step": 1062
+    },
+    {
+      "epoch": 1.3042944785276074,
+      "grad_norm": 9.167928996893854,
+      "learning_rate": 4.5685950413223135e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0537,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.5,
+      "step": 1063
+    },
+    {
+      "epoch": 1.305521472392638,
+      "grad_norm": 8.302137097278727,
+      "learning_rate": 4.5652892561983467e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0359,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.82421875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.0,
+      "step": 1064
+    },
+    {
+      "epoch": 1.3067484662576687,
+      "grad_norm": 10.525223064527983,
+      "learning_rate": 4.56198347107438e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -442.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0462,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7578125,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -9.0625,
+      "step": 1065
+    },
+    {
+      "epoch": 1.3079754601226994,
+      "grad_norm": 10.97929802196726,
+      "learning_rate": 4.558677685950413e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -456.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0383,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.3125,
+      "step": 1066
+    },
+    {
+      "epoch": 1.30920245398773,
+      "grad_norm": 10.890729674094034,
+      "learning_rate": 4.555371900826446e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -364.0,
+      "loss": 0.051,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.49609375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.0,
+      "step": 1067
+    },
+    {
+      "epoch": 1.3104294478527607,
+      "grad_norm": 13.051811444762572,
+      "learning_rate": 4.5520661157024794e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -340.0,
+      "loss": 0.062,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.5,
+      "step": 1068
+    },
+    {
+      "epoch": 1.3116564417177914,
+      "grad_norm": 11.077880050653722,
+      "learning_rate": 4.5487603305785126e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0566,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.625,
+      "step": 1069
+    },
+    {
+      "epoch": 1.312883435582822,
+      "grad_norm": 11.319123638531106,
+      "learning_rate": 4.5454545454545457e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0333,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.5625,
+      "step": 1070
+    },
+    {
+      "epoch": 1.3141104294478527,
+      "grad_norm": 15.58096616127429,
+      "learning_rate": 4.542148760330578e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0632,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.0,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.3125,
+      "step": 1071
+    },
+    {
+      "epoch": 1.3153374233128834,
+      "grad_norm": 9.810761244120958,
+      "learning_rate": 4.538842975206611e-07,
+      "logits/chosen": -0.1201171875,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0379,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -7.8125,
+      "step": 1072
+    },
+    {
+      "epoch": 1.316564417177914,
+      "grad_norm": 13.233835603974985,
+      "learning_rate": 4.535537190082644e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -324.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0544,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.90625,
+      "step": 1073
+    },
+    {
+      "epoch": 1.3177914110429447,
+      "grad_norm": 7.47915926353906,
+      "learning_rate": 4.5322314049586774e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0321,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.4375,
+      "step": 1074
+    },
+    {
+      "epoch": 1.3190184049079754,
+      "grad_norm": 11.99553982285267,
+      "learning_rate": 4.5289256198347106e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0402,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.125,
+      "step": 1075
+    },
+    {
+      "epoch": 1.320245398773006,
+      "grad_norm": 9.482049024318554,
+      "learning_rate": 4.5256198347107437e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0406,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.5,
+      "step": 1076
+    },
+    {
+      "epoch": 1.3214723926380367,
+      "grad_norm": 9.407916883456961,
+      "learning_rate": 4.522314049586777e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -294.0,
+      "logps/rejected": -290.0,
+      "loss": 0.0463,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.390625,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -8.125,
+      "step": 1077
+    },
+    {
+      "epoch": 1.3226993865030674,
+      "grad_norm": 12.437035820876217,
+      "learning_rate": 4.51900826446281e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.44921875,
+      "rewards/margins": 7.84375,
+      "rewards/rejected": -7.40625,
+      "step": 1078
+    },
+    {
+      "epoch": 1.323926380368098,
+      "grad_norm": 13.80505764294533,
+      "learning_rate": 4.515702479338843e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0704,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.4375,
+      "step": 1079
+    },
+    {
+      "epoch": 1.3251533742331287,
+      "grad_norm": 12.147288084424153,
+      "learning_rate": 4.5123966942148754e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0568,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.75,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -7.4375,
+      "step": 1080
+    },
+    {
+      "epoch": 1.3263803680981594,
+      "grad_norm": 7.917444989605004,
+      "learning_rate": 4.5090909090909086e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0379,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.89453125,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.1875,
+      "step": 1081
+    },
+    {
+      "epoch": 1.32760736196319,
+      "grad_norm": 8.384496689172522,
+      "learning_rate": 4.505785123966942e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0376,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -7.6875,
+      "step": 1082
+    },
+    {
+      "epoch": 1.3288343558282207,
+      "grad_norm": 9.161583572655749,
+      "learning_rate": 4.502479338842975e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -306.0,
+      "loss": 0.043,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.5390625,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -7.625,
+      "step": 1083
+    },
+    {
+      "epoch": 1.3300613496932514,
+      "grad_norm": 18.270345545412372,
+      "learning_rate": 4.499173553719008e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0751,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.78125,
+      "step": 1084
+    },
+    {
+      "epoch": 1.331288343558282,
+      "grad_norm": 13.799249791090745,
+      "learning_rate": 4.4958677685950413e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.404296875,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -7.8125,
+      "step": 1085
+    },
+    {
+      "epoch": 1.332515337423313,
+      "grad_norm": 8.946002080660248,
+      "learning_rate": 4.4925619834710744e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -310.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0496,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.25,
+      "step": 1086
+    },
+    {
+      "epoch": 1.3337423312883436,
+      "grad_norm": 10.788580914750392,
+      "learning_rate": 4.4892561983471076e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -318.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1943359375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.3125,
+      "step": 1087
+    },
+    {
+      "epoch": 1.3349693251533743,
+      "grad_norm": 14.32023199723315,
+      "learning_rate": 4.485950413223141e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0476,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.3125,
+      "step": 1088
+    },
+    {
+      "epoch": 1.336196319018405,
+      "grad_norm": 11.907439115072256,
+      "learning_rate": 4.482644628099173e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0615,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.0673828125,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -8.25,
+      "step": 1089
+    },
+    {
+      "epoch": 1.3374233128834356,
+      "grad_norm": 12.503556120707168,
+      "learning_rate": 4.479338842975206e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0519,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.197265625,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -8.0625,
+      "step": 1090
+    },
+    {
+      "epoch": 1.3386503067484663,
+      "grad_norm": 18.17820402883285,
+      "learning_rate": 4.4760330578512393e-07,
+      "logits/chosen": -0.0947265625,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0573,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -7.71875,
+      "step": 1091
+    },
+    {
+      "epoch": 1.339877300613497,
+      "grad_norm": 12.556890306296957,
+      "learning_rate": 4.4727272727272724e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.0625,
+      "step": 1092
+    },
+    {
+      "epoch": 1.3411042944785276,
+      "grad_norm": 10.170118448319533,
+      "learning_rate": 4.4694214876033056e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0468,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7265625,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.3125,
+      "step": 1093
+    },
+    {
+      "epoch": 1.3423312883435583,
+      "grad_norm": 14.331382863628885,
+      "learning_rate": 4.466115702479339e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -400.0,
+      "loss": 0.052,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.4375,
+      "step": 1094
+    },
+    {
+      "epoch": 1.343558282208589,
+      "grad_norm": 8.788938438317068,
+      "learning_rate": 4.462809917355372e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -440.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0296,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -9.375,
+      "step": 1095
+    },
+    {
+      "epoch": 1.3447852760736196,
+      "grad_norm": 22.147235958677946,
+      "learning_rate": 4.459504132231405e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0749,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.1875,
+      "step": 1096
+    },
+    {
+      "epoch": 1.3460122699386503,
+      "grad_norm": 8.228501370409944,
+      "learning_rate": 4.4561983471074383e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -426.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0324,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.375,
+      "step": 1097
+    },
+    {
+      "epoch": 1.347239263803681,
+      "grad_norm": 10.392801869852612,
+      "learning_rate": 4.4528925619834705e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.671875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.0,
+      "step": 1098
+    },
+    {
+      "epoch": 1.3484662576687116,
+      "grad_norm": 19.360520432132724,
+      "learning_rate": 4.4495867768595036e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0913,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.322265625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.1875,
+      "step": 1099
+    },
+    {
+      "epoch": 1.3496932515337423,
+      "grad_norm": 10.418818451409532,
+      "learning_rate": 4.446280991735537e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.490234375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0385,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -8.5625,
+      "step": 1100
+    },
+    {
+      "epoch": 1.350920245398773,
+      "grad_norm": 12.995574614197842,
+      "learning_rate": 4.44297520661157e-07,
+      "logits/chosen": -0.08203125,
+      "logits/rejected": -0.2392578125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0568,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.0,
+      "step": 1101
+    },
+    {
+      "epoch": 1.3521472392638036,
+      "grad_norm": 12.498300123184615,
+      "learning_rate": 4.439669421487603e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0541,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -9.875,
+      "step": 1102
+    },
+    {
+      "epoch": 1.3533742331288343,
+      "grad_norm": 15.391881436542993,
+      "learning_rate": 4.4363636363636363e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0679,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.796875,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.1875,
+      "step": 1103
+    },
+    {
+      "epoch": 1.354601226993865,
+      "grad_norm": 14.536968024679194,
+      "learning_rate": 4.4330578512396695e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0651,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.51953125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.0,
+      "step": 1104
+    },
+    {
+      "epoch": 1.3558282208588956,
+      "grad_norm": 6.1991402014729875,
+      "learning_rate": 4.4297520661157027e-07,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0267,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -9.6875,
+      "step": 1105
+    },
+    {
+      "epoch": 1.3570552147239263,
+      "grad_norm": 9.510072203259035,
+      "learning_rate": 4.426446280991735e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0466,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1357421875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.5,
+      "step": 1106
+    },
+    {
+      "epoch": 1.358282208588957,
+      "grad_norm": 11.345500016334285,
+      "learning_rate": 4.423140495867768e-07,
+      "logits/chosen": -0.0986328125,
+      "logits/rejected": -0.21875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -360.0,
+      "loss": 0.045,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.578125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.1875,
+      "step": 1107
+    },
+    {
+      "epoch": 1.3595092024539877,
+      "grad_norm": 8.67082878371813,
+      "learning_rate": 4.419834710743801e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -442.0,
+      "logps/rejected": -388.0,
+      "loss": 0.035,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.4375,
+      "step": 1108
+    },
+    {
+      "epoch": 1.3607361963190185,
+      "grad_norm": 11.14953002163676,
+      "learning_rate": 4.4165289256198343e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0551,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.6015625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.75,
+      "step": 1109
+    },
+    {
+      "epoch": 1.3619631901840492,
+      "grad_norm": 15.6793165244425,
+      "learning_rate": 4.4132231404958675e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0635,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.75,
+      "step": 1110
+    },
+    {
+      "epoch": 1.3631901840490799,
+      "grad_norm": 7.626396120756006,
+      "learning_rate": 4.4099173553719007e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -434.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0359,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.470703125,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.25,
+      "step": 1111
+    },
+    {
+      "epoch": 1.3644171779141105,
+      "grad_norm": 10.029174865101465,
+      "learning_rate": 4.406611570247934e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0503,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.46484375,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.5,
+      "step": 1112
+    },
+    {
+      "epoch": 1.3656441717791412,
+      "grad_norm": 10.857796592085158,
+      "learning_rate": 4.403305785123967e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0512,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.5,
+      "step": 1113
+    },
+    {
+      "epoch": 1.3668711656441719,
+      "grad_norm": 15.765671968754052,
+      "learning_rate": 4.4e-07,
+      "logits/chosen": -0.3515625,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -344.0,
+      "loss": 0.056,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.0272216796875,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -9.125,
+      "step": 1114
+    },
+    {
+      "epoch": 1.3680981595092025,
+      "grad_norm": 11.976634672452079,
+      "learning_rate": 4.3966942148760323e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0489,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.83984375,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.0625,
+      "step": 1115
+    },
+    {
+      "epoch": 1.3693251533742332,
+      "grad_norm": 12.12629302690259,
+      "learning_rate": 4.3933884297520655e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0589,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.5703125,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.1875,
+      "step": 1116
+    },
+    {
+      "epoch": 1.3705521472392639,
+      "grad_norm": 9.786096050833946,
+      "learning_rate": 4.3900826446280987e-07,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0454,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.9375,
+      "step": 1117
+    },
+    {
+      "epoch": 1.3717791411042946,
+      "grad_norm": 12.232079190104093,
+      "learning_rate": 4.386776859504132e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0636,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.3671875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.6875,
+      "step": 1118
+    },
+    {
+      "epoch": 1.3730061349693252,
+      "grad_norm": 9.928765564322998,
+      "learning_rate": 4.383471074380165e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0526,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.4140625,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -8.0,
+      "step": 1119
+    },
+    {
+      "epoch": 1.3742331288343559,
+      "grad_norm": 10.258302380832172,
+      "learning_rate": 4.380165289256198e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0325,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.625,
+      "step": 1120
+    },
+    {
+      "epoch": 1.3754601226993866,
+      "grad_norm": 20.979255833463608,
+      "learning_rate": 4.3768595041322314e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0915,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.125,
+      "step": 1121
+    },
+    {
+      "epoch": 1.3766871165644172,
+      "grad_norm": 11.325879983054083,
+      "learning_rate": 4.3735537190082646e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0423,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2890625,
+      "rewards/margins": 8.3125,
+      "rewards/rejected": -8.0625,
+      "step": 1122
+    },
+    {
+      "epoch": 1.377914110429448,
+      "grad_norm": 10.020287970518874,
+      "learning_rate": 4.370247933884298e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -436.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0387,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.3515625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.1875,
+      "step": 1123
+    },
+    {
+      "epoch": 1.3791411042944786,
+      "grad_norm": 11.809777379798582,
+      "learning_rate": 4.36694214876033e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0466,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.56640625,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.4375,
+      "step": 1124
+    },
+    {
+      "epoch": 1.3803680981595092,
+      "grad_norm": 8.666119379064078,
+      "learning_rate": 4.363636363636363e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0368,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.484375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.1875,
+      "step": 1125
+    },
+    {
+      "epoch": 1.38159509202454,
+      "grad_norm": 9.38301821790362,
+      "learning_rate": 4.360330578512396e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0395,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.51953125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.375,
+      "step": 1126
+    },
+    {
+      "epoch": 1.3828220858895706,
+      "grad_norm": 13.160048641339023,
+      "learning_rate": 4.3570247933884294e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -442.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0599,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.90625,
+      "step": 1127
+    },
+    {
+      "epoch": 1.3840490797546012,
+      "grad_norm": 14.649656027167724,
+      "learning_rate": 4.3537190082644626e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0755,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.3203125,
+      "rewards/margins": 7.8125,
+      "rewards/rejected": -7.5,
+      "step": 1128
+    },
+    {
+      "epoch": 1.385276073619632,
+      "grad_norm": 8.607132085557556,
+      "learning_rate": 4.350413223140496e-07,
+      "logits/chosen": -0.11376953125,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0342,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.59375,
+      "step": 1129
+    },
+    {
+      "epoch": 1.3865030674846626,
+      "grad_norm": 18.504640307311753,
+      "learning_rate": 4.347107438016529e-07,
+      "logits/chosen": -0.0634765625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0563,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -7.75,
+      "step": 1130
+    },
+    {
+      "epoch": 1.3877300613496932,
+      "grad_norm": 9.275241570046726,
+      "learning_rate": 4.343801652892562e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -7.75,
+      "step": 1131
+    },
+    {
+      "epoch": 1.388957055214724,
+      "grad_norm": 13.545233665904872,
+      "learning_rate": 4.3404958677685953e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -484.0,
+      "logps/rejected": -380.0,
+      "loss": 0.042,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -7.46875,
+      "step": 1132
+    },
+    {
+      "epoch": 1.3901840490797546,
+      "grad_norm": 9.02011133210679,
+      "learning_rate": 4.3371900826446274e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0402,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.0,
+      "step": 1133
+    },
+    {
+      "epoch": 1.3914110429447852,
+      "grad_norm": 9.626052329267141,
+      "learning_rate": 4.3338842975206606e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -316.0,
+      "logps/rejected": -312.0,
+      "loss": 0.0531,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.0732421875,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -8.1875,
+      "step": 1134
+    },
+    {
+      "epoch": 1.392638036809816,
+      "grad_norm": 9.379296233151626,
+      "learning_rate": 4.330578512396694e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.0,
+      "step": 1135
+    },
+    {
+      "epoch": 1.3938650306748466,
+      "grad_norm": 18.617309825421266,
+      "learning_rate": 4.327272727272727e-07,
+      "logits/chosen": -0.357421875,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -458.0,
+      "logps/rejected": -380.0,
+      "loss": 0.067,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.9375,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -8.1875,
+      "step": 1136
+    },
+    {
+      "epoch": 1.3950920245398772,
+      "grad_norm": 16.539416510452526,
+      "learning_rate": 4.32396694214876e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.55078125,
+      "logps/chosen": -458.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0659,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.6875,
+      "step": 1137
+    },
+    {
+      "epoch": 1.396319018404908,
+      "grad_norm": 9.959030275396955,
+      "learning_rate": 4.3206611570247933e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0397,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 8.0,
+      "rewards/rejected": -7.0625,
+      "step": 1138
+    },
+    {
+      "epoch": 1.3975460122699386,
+      "grad_norm": 10.991054262003571,
+      "learning_rate": 4.3173553719008265e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.490234375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0449,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -7.9375,
+      "step": 1139
+    },
+    {
+      "epoch": 1.3987730061349692,
+      "grad_norm": 13.764692227891263,
+      "learning_rate": 4.3140495867768597e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -342.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0488,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.72265625,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.96875,
+      "step": 1140
+    },
+    {
+      "epoch": 1.4,
+      "grad_norm": 9.018367038711595,
+      "learning_rate": 4.310743801652893e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -332.0,
+      "loss": 0.032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.828125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.84375,
+      "step": 1141
+    },
+    {
+      "epoch": 1.4012269938650306,
+      "grad_norm": 14.08834337812807,
+      "learning_rate": 4.307438016528925e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0521,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.875,
+      "step": 1142
+    },
+    {
+      "epoch": 1.4024539877300612,
+      "grad_norm": 10.911725420548523,
+      "learning_rate": 4.304132231404958e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0453,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.515625,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.3125,
+      "step": 1143
+    },
+    {
+      "epoch": 1.403680981595092,
+      "grad_norm": 10.875646708704325,
+      "learning_rate": 4.3008264462809913e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0366,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.89453125,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.1875,
+      "step": 1144
+    },
+    {
+      "epoch": 1.4049079754601226,
+      "grad_norm": 19.771180533625117,
+      "learning_rate": 4.2975206611570245e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -452.0,
+      "logps/rejected": -402.0,
+      "loss": 0.055,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -7.84375,
+      "step": 1145
+    },
+    {
+      "epoch": 1.4061349693251532,
+      "grad_norm": 11.6301527103396,
+      "learning_rate": 4.2942148760330577e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0461,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.4375,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.625,
+      "step": 1146
+    },
+    {
+      "epoch": 1.407361963190184,
+      "grad_norm": 4.878005421347195,
+      "learning_rate": 4.290909090909091e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0252,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.34375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.6875,
+      "step": 1147
+    },
+    {
+      "epoch": 1.4085889570552146,
+      "grad_norm": 8.16557883584569,
+      "learning_rate": 4.287603305785124e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0307,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.5,
+      "step": 1148
+    },
+    {
+      "epoch": 1.4098159509202455,
+      "grad_norm": 8.164045264121068,
+      "learning_rate": 4.284297520661157e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0278,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.9375,
+      "step": 1149
+    },
+    {
+      "epoch": 1.4110429447852761,
+      "grad_norm": 10.04611876987732,
+      "learning_rate": 4.2809917355371904e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0408,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.462890625,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.625,
+      "step": 1150
+    },
+    {
+      "epoch": 1.4122699386503068,
+      "grad_norm": 10.11694907782239,
+      "learning_rate": 4.2776859504132225e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0399,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.482421875,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.875,
+      "step": 1151
+    },
+    {
+      "epoch": 1.4134969325153375,
+      "grad_norm": 6.802420223259662,
+      "learning_rate": 4.2743801652892557e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0252,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.0,
+      "step": 1152
+    },
+    {
+      "epoch": 1.4147239263803681,
+      "grad_norm": 12.518162366564754,
+      "learning_rate": 4.271074380165289e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0518,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -8.8125,
+      "step": 1153
+    },
+    {
+      "epoch": 1.4159509202453988,
+      "grad_norm": 14.27379900942779,
+      "learning_rate": 4.267768595041322e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.470703125,
+      "logps/chosen": -442.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0425,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.1875,
+      "step": 1154
+    },
+    {
+      "epoch": 1.4171779141104295,
+      "grad_norm": 10.309850585087835,
+      "learning_rate": 4.264462809917355e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0442,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1865234375,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.8125,
+      "step": 1155
+    },
+    {
+      "epoch": 1.4184049079754601,
+      "grad_norm": 15.590234943117638,
+      "learning_rate": 4.2611570247933884e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -434.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0428,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -9.5,
+      "step": 1156
+    },
+    {
+      "epoch": 1.4196319018404908,
+      "grad_norm": 9.987550517532815,
+      "learning_rate": 4.2578512396694215e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -446.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0317,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -8.9375,
+      "step": 1157
+    },
+    {
+      "epoch": 1.4208588957055215,
+      "grad_norm": 7.931416115322814,
+      "learning_rate": 4.2545454545454547e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -342.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0371,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.828125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.0625,
+      "step": 1158
+    },
+    {
+      "epoch": 1.4220858895705522,
+      "grad_norm": 18.05171696034389,
+      "learning_rate": 4.251239669421488e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0767,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.296875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.8125,
+      "step": 1159
+    },
+    {
+      "epoch": 1.4233128834355828,
+      "grad_norm": 13.943376246751296,
+      "learning_rate": 4.24793388429752e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0449,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.65625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.875,
+      "step": 1160
+    },
+    {
+      "epoch": 1.4245398773006135,
+      "grad_norm": 12.317132054833733,
+      "learning_rate": 4.244628099173553e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0585,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.75,
+      "step": 1161
+    },
+    {
+      "epoch": 1.4257668711656442,
+      "grad_norm": 9.354899646749724,
+      "learning_rate": 4.2413223140495864e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0414,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.0,
+      "step": 1162
+    },
+    {
+      "epoch": 1.4269938650306748,
+      "grad_norm": 8.354103036645931,
+      "learning_rate": 4.2380165289256195e-07,
+      "logits/chosen": -0.353515625,
+      "logits/rejected": -0.59765625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0303,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.4375,
+      "step": 1163
+    },
+    {
+      "epoch": 1.4282208588957055,
+      "grad_norm": 17.92923732721561,
+      "learning_rate": 4.2347107438016527e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.24609375,
+      "logps/chosen": -330.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0599,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.53125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.75,
+      "step": 1164
+    },
+    {
+      "epoch": 1.4294478527607362,
+      "grad_norm": 8.92045181790276,
+      "learning_rate": 4.231404958677686e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0279,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.3671875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.125,
+      "step": 1165
+    },
+    {
+      "epoch": 1.4306748466257668,
+      "grad_norm": 9.823343955782265,
+      "learning_rate": 4.228099173553719e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0317,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.5625,
+      "step": 1166
+    },
+    {
+      "epoch": 1.4319018404907975,
+      "grad_norm": 8.303351669646478,
+      "learning_rate": 4.224793388429752e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0256,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.6328125,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.6875,
+      "step": 1167
+    },
+    {
+      "epoch": 1.4331288343558282,
+      "grad_norm": 13.096682943463716,
+      "learning_rate": 4.2214876033057854e-07,
+      "logits/chosen": -0.248046875,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0432,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.3359375,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.8125,
+      "step": 1168
+    },
+    {
+      "epoch": 1.4343558282208588,
+      "grad_norm": 8.383755609382192,
+      "learning_rate": 4.2181818181818175e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0339,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.71875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.25,
+      "step": 1169
+    },
+    {
+      "epoch": 1.4355828220858895,
+      "grad_norm": 13.087573566535397,
+      "learning_rate": 4.2148760330578507e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0444,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.443359375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.1875,
+      "step": 1170
+    },
+    {
+      "epoch": 1.4368098159509202,
+      "grad_norm": 11.843632840671855,
+      "learning_rate": 4.211570247933884e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0424,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -8.9375,
+      "step": 1171
+    },
+    {
+      "epoch": 1.438036809815951,
+      "grad_norm": 13.430060643840084,
+      "learning_rate": 4.208264462809917e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0613,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.287109375,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -8.0625,
+      "step": 1172
+    },
+    {
+      "epoch": 1.4392638036809817,
+      "grad_norm": 13.140000801988304,
+      "learning_rate": 4.20495867768595e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0435,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.61328125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.9375,
+      "step": 1173
+    },
+    {
+      "epoch": 1.4404907975460124,
+      "grad_norm": 10.037701828497555,
+      "learning_rate": 4.2016528925619834e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0387,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.3125,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.75,
+      "step": 1174
+    },
+    {
+      "epoch": 1.441717791411043,
+      "grad_norm": 7.925062400196428,
+      "learning_rate": 4.1983471074380166e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -342.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0409,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.12890625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.75,
+      "step": 1175
+    },
+    {
+      "epoch": 1.4429447852760737,
+      "grad_norm": 7.918415538616964,
+      "learning_rate": 4.19504132231405e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0375,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1015625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.5,
+      "step": 1176
+    },
+    {
+      "epoch": 1.4441717791411044,
+      "grad_norm": 9.953622472970856,
+      "learning_rate": 4.191735537190083e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0429,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.0,
+      "step": 1177
+    },
+    {
+      "epoch": 1.445398773006135,
+      "grad_norm": 9.134426899002403,
+      "learning_rate": 4.188429752066115e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0339,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -8.5,
+      "step": 1178
+    },
+    {
+      "epoch": 1.4466257668711657,
+      "grad_norm": 9.094305897866864,
+      "learning_rate": 4.185123966942148e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0398,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.625,
+      "step": 1179
+    },
+    {
+      "epoch": 1.4478527607361964,
+      "grad_norm": 13.578461286413251,
+      "learning_rate": 4.1818181818181814e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0506,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.46484375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.875,
+      "step": 1180
+    },
+    {
+      "epoch": 1.449079754601227,
+      "grad_norm": 7.992882025383505,
+      "learning_rate": 4.1785123966942146e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0253,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.9375,
+      "step": 1181
+    },
+    {
+      "epoch": 1.4503067484662577,
+      "grad_norm": 10.0106751323216,
+      "learning_rate": 4.175206611570248e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -384.0,
+      "loss": 0.036,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.21875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.625,
+      "step": 1182
+    },
+    {
+      "epoch": 1.4515337423312884,
+      "grad_norm": 15.000115439366018,
+      "learning_rate": 4.171900826446281e-07,
+      "logits/chosen": -0.24609375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0654,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.193359375,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -8.0,
+      "step": 1183
+    },
+    {
+      "epoch": 1.452760736196319,
+      "grad_norm": 10.305727034916446,
+      "learning_rate": 4.168595041322314e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0461,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.72265625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.8125,
+      "step": 1184
+    },
+    {
+      "epoch": 1.4539877300613497,
+      "grad_norm": 10.606454430153265,
+      "learning_rate": 4.1652892561983473e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0407,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.3125,
+      "step": 1185
+    },
+    {
+      "epoch": 1.4552147239263804,
+      "grad_norm": 9.636022481666009,
+      "learning_rate": 4.1619834710743805e-07,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0559,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.96875,
+      "step": 1186
+    },
+    {
+      "epoch": 1.456441717791411,
+      "grad_norm": 11.476673812202115,
+      "learning_rate": 4.1586776859504126e-07,
+      "logits/chosen": -0.2421875,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -450.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0466,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -7.625,
+      "step": 1187
+    },
+    {
+      "epoch": 1.4576687116564417,
+      "grad_norm": 11.105210699415723,
+      "learning_rate": 4.155371900826446e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.9375,
+      "step": 1188
+    },
+    {
+      "epoch": 1.4588957055214724,
+      "grad_norm": 12.661874309989074,
+      "learning_rate": 4.152066115702479e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0575,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.125,
+      "step": 1189
+    },
+    {
+      "epoch": 1.460122699386503,
+      "grad_norm": 8.654978193473417,
+      "learning_rate": 4.148760330578512e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0357,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.5625,
+      "step": 1190
+    },
+    {
+      "epoch": 1.4613496932515337,
+      "grad_norm": 11.584776568508845,
+      "learning_rate": 4.1454545454545453e-07,
+      "logits/chosen": -0.11376953125,
+      "logits/rejected": -0.25,
+      "logps/chosen": -364.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0451,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.75,
+      "step": 1191
+    },
+    {
+      "epoch": 1.4625766871165644,
+      "grad_norm": 11.938546188570182,
+      "learning_rate": 4.1421487603305785e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0479,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.75,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.375,
+      "step": 1192
+    },
+    {
+      "epoch": 1.463803680981595,
+      "grad_norm": 8.57539996640767,
+      "learning_rate": 4.1388429752066117e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -452.0,
+      "logps/rejected": -390.0,
+      "loss": 0.029,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.0625,
+      "step": 1193
+    },
+    {
+      "epoch": 1.4650306748466257,
+      "grad_norm": 9.189664993288334,
+      "learning_rate": 4.135537190082645e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -382.0,
+      "loss": 0.021,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.375,
+      "step": 1194
+    },
+    {
+      "epoch": 1.4662576687116564,
+      "grad_norm": 9.081781667653363,
+      "learning_rate": 4.132231404958677e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -384.0,
+      "loss": 0.04,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.76171875,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.25,
+      "step": 1195
+    },
+    {
+      "epoch": 1.467484662576687,
+      "grad_norm": 12.415433200199677,
+      "learning_rate": 4.12892561983471e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0524,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.4453125,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.96875,
+      "step": 1196
+    },
+    {
+      "epoch": 1.4687116564417177,
+      "grad_norm": 9.943422765866982,
+      "learning_rate": 4.1256198347107433e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0466,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.25,
+      "step": 1197
+    },
+    {
+      "epoch": 1.4699386503067484,
+      "grad_norm": 14.559683333211177,
+      "learning_rate": 4.1223140495867765e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0527,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.73828125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.96875,
+      "step": 1198
+    },
+    {
+      "epoch": 1.471165644171779,
+      "grad_norm": 17.776993736511606,
+      "learning_rate": 4.1190082644628097e-07,
+      "logits/chosen": -0.31640625,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0515,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.6875,
+      "step": 1199
+    },
+    {
+      "epoch": 1.4723926380368098,
+      "grad_norm": 7.694268373865595,
+      "learning_rate": 4.115702479338843e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0276,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.55859375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.9375,
+      "step": 1200
+    },
+    {
+      "epoch": 1.4736196319018404,
+      "grad_norm": 7.758540238755445,
+      "learning_rate": 4.112396694214876e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0271,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.25,
+      "step": 1201
+    },
+    {
+      "epoch": 1.474846625766871,
+      "grad_norm": 9.980200405050743,
+      "learning_rate": 4.109090909090909e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0357,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.375,
+      "step": 1202
+    },
+    {
+      "epoch": 1.4760736196319018,
+      "grad_norm": 8.852224972647532,
+      "learning_rate": 4.1057851239669424e-07,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0434,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.66015625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.4375,
+      "step": 1203
+    },
+    {
+      "epoch": 1.4773006134969324,
+      "grad_norm": 10.918645848118,
+      "learning_rate": 4.1024793388429745e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0408,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.4375,
+      "step": 1204
+    },
+    {
+      "epoch": 1.478527607361963,
+      "grad_norm": 15.360319822899514,
+      "learning_rate": 4.0991735537190077e-07,
+      "logits/chosen": -0.1064453125,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0614,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.1875,
+      "step": 1205
+    },
+    {
+      "epoch": 1.4797546012269938,
+      "grad_norm": 6.938456095237597,
+      "learning_rate": 4.095867768595041e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -344.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0355,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.55078125,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.3125,
+      "step": 1206
+    },
+    {
+      "epoch": 1.4809815950920244,
+      "grad_norm": 12.838247596989346,
+      "learning_rate": 4.092561983471074e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0441,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.296875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.5,
+      "step": 1207
+    },
+    {
+      "epoch": 1.482208588957055,
+      "grad_norm": 14.774384990299971,
+      "learning_rate": 4.089256198347107e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0675,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.294921875,
+      "rewards/margins": 8.3125,
+      "rewards/rejected": -8.0625,
+      "step": 1208
+    },
+    {
+      "epoch": 1.4834355828220858,
+      "grad_norm": 11.62853808692027,
+      "learning_rate": 4.0859504132231404e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0536,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.47265625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.875,
+      "step": 1209
+    },
+    {
+      "epoch": 1.4846625766871164,
+      "grad_norm": 11.857036692599188,
+      "learning_rate": 4.0826446280991736e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0305,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.85546875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.5625,
+      "step": 1210
+    },
+    {
+      "epoch": 1.4858895705521473,
+      "grad_norm": 4.69235230102138,
+      "learning_rate": 4.079338842975207e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -360.0,
+      "loss": 0.019,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.625,
+      "step": 1211
+    },
+    {
+      "epoch": 1.487116564417178,
+      "grad_norm": 9.185641508263624,
+      "learning_rate": 4.07603305785124e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0428,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.326171875,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.9375,
+      "step": 1212
+    },
+    {
+      "epoch": 1.4883435582822087,
+      "grad_norm": 8.195477892364691,
+      "learning_rate": 4.072727272727272e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.55078125,
+      "logps/chosen": -428.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0365,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.75,
+      "step": 1213
+    },
+    {
+      "epoch": 1.4895705521472393,
+      "grad_norm": 14.570501254636081,
+      "learning_rate": 4.069421487603305e-07,
+      "logits/chosen": -0.2421875,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -346.0,
+      "logps/rejected": -356.0,
+      "loss": 0.056,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": -0.018310546875,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -9.0625,
+      "step": 1214
+    },
+    {
+      "epoch": 1.49079754601227,
+      "grad_norm": 11.800836028244067,
+      "learning_rate": 4.0661157024793384e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -378.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0324,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.5625,
+      "step": 1215
+    },
+    {
+      "epoch": 1.4920245398773007,
+      "grad_norm": 12.335957050020774,
+      "learning_rate": 4.0628099173553716e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0482,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.00836181640625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.625,
+      "step": 1216
+    },
+    {
+      "epoch": 1.4932515337423313,
+      "grad_norm": 9.006796955903845,
+      "learning_rate": 4.059504132231405e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -382.0,
+      "loss": 0.03,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.6875,
+      "step": 1217
+    },
+    {
+      "epoch": 1.494478527607362,
+      "grad_norm": 8.123325606423187,
+      "learning_rate": 4.056198347107438e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0321,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.01385498046875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.9375,
+      "step": 1218
+    },
+    {
+      "epoch": 1.4957055214723927,
+      "grad_norm": 12.558099415012947,
+      "learning_rate": 4.052892561983471e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -388.0,
+      "loss": 0.036,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.4296875,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.8125,
+      "step": 1219
+    },
+    {
+      "epoch": 1.4969325153374233,
+      "grad_norm": 12.775377284659152,
+      "learning_rate": 4.0495867768595043e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0506,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.2734375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.875,
+      "step": 1220
+    },
+    {
+      "epoch": 1.498159509202454,
+      "grad_norm": 11.306165820615583,
+      "learning_rate": 4.0462809917355375e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0388,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.71875,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.125,
+      "step": 1221
+    },
+    {
+      "epoch": 1.4993865030674847,
+      "grad_norm": 12.268638749924271,
+      "learning_rate": 4.0429752066115696e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -334.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.33984375,
+      "rewards/margins": 8.3125,
+      "rewards/rejected": -8.625,
+      "step": 1222
+    },
+    {
+      "epoch": 1.5006134969325153,
+      "grad_norm": 16.447342830994092,
+      "learning_rate": 4.039669421487603e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -378.0,
+      "loss": 0.049,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.5625,
+      "step": 1223
+    },
+    {
+      "epoch": 1.501840490797546,
+      "grad_norm": 10.490780921280326,
+      "learning_rate": 4.036363636363636e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0536,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.458984375,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -8.4375,
+      "step": 1224
+    },
+    {
+      "epoch": 1.5030674846625767,
+      "grad_norm": 8.180288581649103,
+      "learning_rate": 4.033057851239669e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0337,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.33203125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.5625,
+      "step": 1225
+    },
+    {
+      "epoch": 1.5042944785276073,
+      "grad_norm": 11.757770803193841,
+      "learning_rate": 4.0297520661157023e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -358.0,
+      "loss": 0.035,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.0179443359375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.6875,
+      "step": 1226
+    },
+    {
+      "epoch": 1.505521472392638,
+      "grad_norm": 7.402348890852564,
+      "learning_rate": 4.0264462809917355e-07,
+      "logits/chosen": -0.38671875,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0259,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.625,
+      "step": 1227
+    },
+    {
+      "epoch": 1.5067484662576687,
+      "grad_norm": 24.868807973648707,
+      "learning_rate": 4.0231404958677686e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -400.0,
+      "loss": 0.075,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.486328125,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.25,
+      "step": 1228
+    },
+    {
+      "epoch": 1.5079754601226993,
+      "grad_norm": 10.416352098408494,
+      "learning_rate": 4.019834710743802e-07,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0321,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.330078125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.6875,
+      "step": 1229
+    },
+    {
+      "epoch": 1.50920245398773,
+      "grad_norm": 12.727906851382768,
+      "learning_rate": 4.016528925619835e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0488,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.74609375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.96875,
+      "step": 1230
+    },
+    {
+      "epoch": 1.510429447852761,
+      "grad_norm": 7.3509385262124125,
+      "learning_rate": 4.013223140495867e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -348.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0284,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.75390625,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -7.78125,
+      "step": 1231
+    },
+    {
+      "epoch": 1.5116564417177916,
+      "grad_norm": 9.498443490772354,
+      "learning_rate": 4.0099173553719003e-07,
+      "logits/chosen": -0.37109375,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0265,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.2109375,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -8.8125,
+      "step": 1232
+    },
+    {
+      "epoch": 1.5128834355828222,
+      "grad_norm": 9.88336081694715,
+      "learning_rate": 4.0066115702479335e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0359,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.375,
+      "step": 1233
+    },
+    {
+      "epoch": 1.514110429447853,
+      "grad_norm": 20.777045013839473,
+      "learning_rate": 4.0033057851239666e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0549,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -7.84375,
+      "step": 1234
+    },
+    {
+      "epoch": 1.5153374233128836,
+      "grad_norm": 11.047074969488666,
+      "learning_rate": 4e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0423,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.171875,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -8.0,
+      "step": 1235
+    },
+    {
+      "epoch": 1.5165644171779142,
+      "grad_norm": 13.010857169907007,
+      "learning_rate": 3.996694214876033e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0398,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.71875,
+      "step": 1236
+    },
+    {
+      "epoch": 1.517791411042945,
+      "grad_norm": 14.316784677028014,
+      "learning_rate": 3.993388429752066e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.5,
+      "logps/chosen": -424.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0461,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.90234375,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.0625,
+      "step": 1237
+    },
+    {
+      "epoch": 1.5190184049079756,
+      "grad_norm": 10.925753336150258,
+      "learning_rate": 3.990082644628099e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -458.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0332,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.125,
+      "step": 1238
+    },
+    {
+      "epoch": 1.5202453987730062,
+      "grad_norm": 14.155768697592602,
+      "learning_rate": 3.986776859504132e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0563,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.9375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.625,
+      "step": 1239
+    },
+    {
+      "epoch": 1.521472392638037,
+      "grad_norm": 12.010081911987827,
+      "learning_rate": 3.983471074380165e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0542,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.68359375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.3125,
+      "step": 1240
+    },
+    {
+      "epoch": 1.5226993865030676,
+      "grad_norm": 15.161747078373114,
+      "learning_rate": 3.9801652892561984e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0635,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.96875,
+      "step": 1241
+    },
+    {
+      "epoch": 1.5239263803680982,
+      "grad_norm": 18.12655990544915,
+      "learning_rate": 3.976859504132231e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0519,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.359375,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.25,
+      "step": 1242
+    },
+    {
+      "epoch": 1.525153374233129,
+      "grad_norm": 12.22688926301596,
+      "learning_rate": 3.973553719008264e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -326.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0418,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.11376953125,
+      "rewards/margins": 8.3125,
+      "rewards/rejected": -8.1875,
+      "step": 1243
+    },
+    {
+      "epoch": 1.5263803680981596,
+      "grad_norm": 7.942928883549814,
+      "learning_rate": 3.9702479338842974e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -336.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0312,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1552734375,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -8.0625,
+      "step": 1244
+    },
+    {
+      "epoch": 1.5276073619631902,
+      "grad_norm": 15.392015274267582,
+      "learning_rate": 3.9669421487603305e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0783,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.6796875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.9375,
+      "step": 1245
+    },
+    {
+      "epoch": 1.528834355828221,
+      "grad_norm": 6.4157878266794315,
+      "learning_rate": 3.9636363636363637e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0324,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.482421875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.25,
+      "step": 1246
+    },
+    {
+      "epoch": 1.5300613496932516,
+      "grad_norm": 6.140011997029994,
+      "learning_rate": 3.9603305785123964e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0233,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.85546875,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.4375,
+      "step": 1247
+    },
+    {
+      "epoch": 1.5312883435582823,
+      "grad_norm": 8.639626779722814,
+      "learning_rate": 3.9570247933884295e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0254,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.8828125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.0,
+      "step": 1248
+    },
+    {
+      "epoch": 1.532515337423313,
+      "grad_norm": 15.793514748193386,
+      "learning_rate": 3.9537190082644627e-07,
+      "logits/chosen": -0.06884765625,
+      "logits/rejected": -0.1806640625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0715,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.09912109375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -9.125,
+      "step": 1249
+    },
+    {
+      "epoch": 1.5337423312883436,
+      "grad_norm": 12.584742135566902,
+      "learning_rate": 3.950413223140496e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -442.0,
+      "logps/rejected": -446.0,
+      "loss": 0.0431,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.81640625,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -10.0625,
+      "step": 1250
+    },
+    {
+      "epoch": 1.5349693251533743,
+      "grad_norm": 16.598890486502036,
+      "learning_rate": 3.9471074380165285e-07,
+      "logits/chosen": -0.37109375,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0551,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -9.6875,
+      "step": 1251
+    },
+    {
+      "epoch": 1.536196319018405,
+      "grad_norm": 14.5035620684537,
+      "learning_rate": 3.9438016528925617e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -352.0,
+      "loss": 0.053,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.3828125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -10.0625,
+      "step": 1252
+    },
+    {
+      "epoch": 1.5374233128834356,
+      "grad_norm": 13.289899357850107,
+      "learning_rate": 3.940495867768595e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0467,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.76953125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.125,
+      "step": 1253
+    },
+    {
+      "epoch": 1.5386503067484663,
+      "grad_norm": 13.876288513389907,
+      "learning_rate": 3.937190082644628e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0566,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.9375,
+      "step": 1254
+    },
+    {
+      "epoch": 1.539877300613497,
+      "grad_norm": 10.631076981710617,
+      "learning_rate": 3.933884297520661e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0369,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.859375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.0625,
+      "step": 1255
+    },
+    {
+      "epoch": 1.5411042944785276,
+      "grad_norm": 13.82590794396477,
+      "learning_rate": 3.930578512396694e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0441,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.59375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.5,
+      "step": 1256
+    },
+    {
+      "epoch": 1.5423312883435583,
+      "grad_norm": 14.32921649123448,
+      "learning_rate": 3.927272727272727e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0511,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.400390625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.875,
+      "step": 1257
+    },
+    {
+      "epoch": 1.543558282208589,
+      "grad_norm": 10.584191334594715,
+      "learning_rate": 3.92396694214876e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0399,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -8.625,
+      "step": 1258
+    },
+    {
+      "epoch": 1.5447852760736196,
+      "grad_norm": 10.748301553600525,
+      "learning_rate": 3.9206611570247934e-07,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0365,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.78515625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.8125,
+      "step": 1259
+    },
+    {
+      "epoch": 1.5460122699386503,
+      "grad_norm": 17.235825888413157,
+      "learning_rate": 3.917355371900826e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -450.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0427,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.9765625,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.9375,
+      "step": 1260
+    },
+    {
+      "epoch": 1.547239263803681,
+      "grad_norm": 10.355163477842352,
+      "learning_rate": 3.914049586776859e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0339,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.56640625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.3125,
+      "step": 1261
+    },
+    {
+      "epoch": 1.5484662576687116,
+      "grad_norm": 11.04210181900335,
+      "learning_rate": 3.9107438016528924e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0509,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.267578125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.6875,
+      "step": 1262
+    },
+    {
+      "epoch": 1.5496932515337423,
+      "grad_norm": 10.926373739741395,
+      "learning_rate": 3.9074380165289256e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0472,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.5625,
+      "step": 1263
+    },
+    {
+      "epoch": 1.550920245398773,
+      "grad_norm": 15.760559572323228,
+      "learning_rate": 3.904132231404958e-07,
+      "logits/chosen": -0.11474609375,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0745,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.453125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.4375,
+      "step": 1264
+    },
+    {
+      "epoch": 1.5521472392638036,
+      "grad_norm": 8.333997789167297,
+      "learning_rate": 3.9008264462809914e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0378,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.51171875,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.75,
+      "step": 1265
+    },
+    {
+      "epoch": 1.5533742331288343,
+      "grad_norm": 11.756550407636311,
+      "learning_rate": 3.8975206611570246e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0417,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.255859375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.4375,
+      "step": 1266
+    },
+    {
+      "epoch": 1.554601226993865,
+      "grad_norm": 16.803054241328326,
+      "learning_rate": 3.894214876033058e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -344.0,
+      "loss": 0.095,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 0.80859375,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.375,
+      "step": 1267
+    },
+    {
+      "epoch": 1.5558282208588956,
+      "grad_norm": 9.056787760649906,
+      "learning_rate": 3.890909090909091e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -496.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0293,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2109375,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.5625,
+      "step": 1268
+    },
+    {
+      "epoch": 1.5570552147239263,
+      "grad_norm": 11.529113923647646,
+      "learning_rate": 3.8876033057851236e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0361,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.4375,
+      "step": 1269
+    },
+    {
+      "epoch": 1.558282208588957,
+      "grad_norm": 12.885170546155125,
+      "learning_rate": 3.884297520661157e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0353,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.67578125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.1875,
+      "step": 1270
+    },
+    {
+      "epoch": 1.5595092024539876,
+      "grad_norm": 9.512525534029331,
+      "learning_rate": 3.88099173553719e-07,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.26953125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0345,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.65625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.25,
+      "step": 1271
+    },
+    {
+      "epoch": 1.5607361963190183,
+      "grad_norm": 5.776653205274608,
+      "learning_rate": 3.877685950413223e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0195,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.3125,
+      "step": 1272
+    },
+    {
+      "epoch": 1.561963190184049,
+      "grad_norm": 18.75333767957364,
+      "learning_rate": 3.874380165289256e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0867,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.625,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.125,
+      "step": 1273
+    },
+    {
+      "epoch": 1.5631901840490796,
+      "grad_norm": 6.671859500637093,
+      "learning_rate": 3.871074380165289e-07,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0226,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.4375,
+      "step": 1274
+    },
+    {
+      "epoch": 1.5644171779141103,
+      "grad_norm": 14.51239466060054,
+      "learning_rate": 3.867768595041322e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0681,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.419921875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.4375,
+      "step": 1275
+    },
+    {
+      "epoch": 1.565644171779141,
+      "grad_norm": 6.983571368573247,
+      "learning_rate": 3.8644628099173553e-07,
+      "logits/chosen": -0.12109375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0252,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.1875,
+      "step": 1276
+    },
+    {
+      "epoch": 1.5668711656441716,
+      "grad_norm": 20.362210341411654,
+      "learning_rate": 3.8611570247933885e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0691,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.431640625,
+      "rewards/margins": 8.25,
+      "rewards/rejected": -7.84375,
+      "step": 1277
+    },
+    {
+      "epoch": 1.5680981595092025,
+      "grad_norm": 9.297990677048228,
+      "learning_rate": 3.857851239669421e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0298,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.4296875,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -8.75,
+      "step": 1278
+    },
+    {
+      "epoch": 1.5693251533742332,
+      "grad_norm": 13.341995153514718,
+      "learning_rate": 3.8545454545454543e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0559,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.6328125,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -7.78125,
+      "step": 1279
+    },
+    {
+      "epoch": 1.5705521472392638,
+      "grad_norm": 12.819101532118133,
+      "learning_rate": 3.8512396694214875e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0446,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.6875,
+      "step": 1280
+    },
+    {
+      "epoch": 1.5717791411042945,
+      "grad_norm": 12.392770855720375,
+      "learning_rate": 3.8479338842975207e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.470703125,
+      "logps/chosen": -432.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0467,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -8.6875,
+      "step": 1281
+    },
+    {
+      "epoch": 1.5730061349693252,
+      "grad_norm": 10.597792635666673,
+      "learning_rate": 3.8446280991735533e-07,
+      "logits/chosen": -0.04541015625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0369,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9140625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.375,
+      "step": 1282
+    },
+    {
+      "epoch": 1.5742331288343558,
+      "grad_norm": 13.234069958439528,
+      "learning_rate": 3.8413223140495865e-07,
+      "logits/chosen": -0.09521484375,
+      "logits/rejected": -0.2275390625,
+      "logps/chosen": -344.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0541,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.08251953125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.8125,
+      "step": 1283
+    },
+    {
+      "epoch": 1.5754601226993865,
+      "grad_norm": 21.414790419464424,
+      "learning_rate": 3.8380165289256197e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0854,
+      "rewards/accuracies": 0.953125,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.1875,
+      "step": 1284
+    },
+    {
+      "epoch": 1.5766871165644172,
+      "grad_norm": 14.621409356775658,
+      "learning_rate": 3.834710743801653e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0331,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.125,
+      "step": 1285
+    },
+    {
+      "epoch": 1.5779141104294478,
+      "grad_norm": 11.226540370477307,
+      "learning_rate": 3.831404958677686e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0366,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.318359375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.0625,
+      "step": 1286
+    },
+    {
+      "epoch": 1.5791411042944785,
+      "grad_norm": 16.039598977424404,
+      "learning_rate": 3.8280991735537187e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -318.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0613,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": -0.1640625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -9.3125,
+      "step": 1287
+    },
+    {
+      "epoch": 1.5803680981595092,
+      "grad_norm": 12.836865540318971,
+      "learning_rate": 3.824793388429752e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0505,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.65625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.0,
+      "step": 1288
+    },
+    {
+      "epoch": 1.5815950920245399,
+      "grad_norm": 7.945536104012517,
+      "learning_rate": 3.821487603305785e-07,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -324.0,
+      "logps/rejected": -318.0,
+      "loss": 0.043,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.072265625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -9.25,
+      "step": 1289
+    },
+    {
+      "epoch": 1.5828220858895705,
+      "grad_norm": 9.269509711129304,
+      "learning_rate": 3.818181818181818e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0298,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1533203125,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -10.125,
+      "step": 1290
+    },
+    {
+      "epoch": 1.5840490797546012,
+      "grad_norm": 12.776471775676947,
+      "learning_rate": 3.814876033057851e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -346.0,
+      "loss": 0.042,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.28125,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.3125,
+      "step": 1291
+    },
+    {
+      "epoch": 1.5852760736196319,
+      "grad_norm": 6.710032759787997,
+      "learning_rate": 3.811570247933884e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0285,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.412109375,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.875,
+      "step": 1292
+    },
+    {
+      "epoch": 1.5865030674846625,
+      "grad_norm": 9.554015835107709,
+      "learning_rate": 3.808264462809917e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0361,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.447265625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.0625,
+      "step": 1293
+    },
+    {
+      "epoch": 1.5877300613496934,
+      "grad_norm": 7.6540954238448995,
+      "learning_rate": 3.8049586776859504e-07,
+      "logits/chosen": -0.2001953125,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0283,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.4296875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.9375,
+      "step": 1294
+    },
+    {
+      "epoch": 1.588957055214724,
+      "grad_norm": 10.249402612807254,
+      "learning_rate": 3.8016528925619836e-07,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0528,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.025390625,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -9.1875,
+      "step": 1295
+    },
+    {
+      "epoch": 1.5901840490797547,
+      "grad_norm": 21.386881079863016,
+      "learning_rate": 3.798347107438016e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -374.0,
+      "loss": 0.1043,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": -0.474609375,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -9.25,
+      "step": 1296
+    },
+    {
+      "epoch": 1.5914110429447854,
+      "grad_norm": 13.13883607900759,
+      "learning_rate": 3.7950413223140494e-07,
+      "logits/chosen": -0.1513671875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0505,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.271484375,
+      "rewards/margins": 8.375,
+      "rewards/rejected": -8.0625,
+      "step": 1297
+    },
+    {
+      "epoch": 1.592638036809816,
+      "grad_norm": 14.148949473846967,
+      "learning_rate": 3.7917355371900826e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0601,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.73828125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.1875,
+      "step": 1298
+    },
+    {
+      "epoch": 1.5938650306748468,
+      "grad_norm": 12.160973266195128,
+      "learning_rate": 3.788429752066116e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0443,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.2353515625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.875,
+      "step": 1299
+    },
+    {
+      "epoch": 1.5950920245398774,
+      "grad_norm": 11.389665455717337,
+      "learning_rate": 3.7851239669421484e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0408,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5546875,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.0,
+      "step": 1300
+    },
+    {
+      "epoch": 1.596319018404908,
+      "grad_norm": 12.77732004217483,
+      "learning_rate": 3.7818181818181816e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0424,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1162109375,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.8125,
+      "step": 1301
+    },
+    {
+      "epoch": 1.5975460122699388,
+      "grad_norm": 12.748510423611386,
+      "learning_rate": 3.778512396694215e-07,
+      "logits/chosen": -0.359375,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0616,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.0,
+      "step": 1302
+    },
+    {
+      "epoch": 1.5987730061349694,
+      "grad_norm": 8.32794148442266,
+      "learning_rate": 3.775206611570248e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0402,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.89453125,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.3125,
+      "step": 1303
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 6.529193782585084,
+      "learning_rate": 3.7719008264462806e-07,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0274,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.84375,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.125,
+      "step": 1304
+    },
+    {
+      "epoch": 1.6012269938650308,
+      "grad_norm": 9.816683371253708,
+      "learning_rate": 3.768595041322314e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -342.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0524,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.4375,
+      "step": 1305
+    },
+    {
+      "epoch": 1.6024539877300614,
+      "grad_norm": 8.33652199166503,
+      "learning_rate": 3.765289256198347e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0333,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.8125,
+      "step": 1306
+    },
+    {
+      "epoch": 1.603680981595092,
+      "grad_norm": 5.703980454598279,
+      "learning_rate": 3.76198347107438e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0237,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.1875,
+      "step": 1307
+    },
+    {
+      "epoch": 1.6049079754601228,
+      "grad_norm": 7.658766807333126,
+      "learning_rate": 3.7586776859504133e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0286,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.9140625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.4375,
+      "step": 1308
+    },
+    {
+      "epoch": 1.6061349693251534,
+      "grad_norm": 9.415605561419792,
+      "learning_rate": 3.755371900826446e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0372,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.50390625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.8125,
+      "step": 1309
+    },
+    {
+      "epoch": 1.607361963190184,
+      "grad_norm": 8.93902350488852,
+      "learning_rate": 3.752066115702479e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -330.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0376,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.375,
+      "step": 1310
+    },
+    {
+      "epoch": 1.6085889570552148,
+      "grad_norm": 12.399527167412383,
+      "learning_rate": 3.7487603305785123e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0605,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.5703125,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -7.59375,
+      "step": 1311
+    },
+    {
+      "epoch": 1.6098159509202454,
+      "grad_norm": 9.954372711363225,
+      "learning_rate": 3.7454545454545454e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0291,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.25,
+      "step": 1312
+    },
+    {
+      "epoch": 1.611042944785276,
+      "grad_norm": 8.770279110550883,
+      "learning_rate": 3.742148760330578e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0342,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.57421875,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.4375,
+      "step": 1313
+    },
+    {
+      "epoch": 1.6122699386503068,
+      "grad_norm": 8.372096906172947,
+      "learning_rate": 3.7388429752066113e-07,
+      "logits/chosen": -0.375,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0239,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.1875,
+      "step": 1314
+    },
+    {
+      "epoch": 1.6134969325153374,
+      "grad_norm": 13.050858951846532,
+      "learning_rate": 3.7355371900826445e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -446.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0497,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9375,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.5625,
+      "step": 1315
+    },
+    {
+      "epoch": 1.614723926380368,
+      "grad_norm": 11.087298553686718,
+      "learning_rate": 3.7322314049586776e-07,
+      "logits/chosen": -0.1103515625,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0421,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.423828125,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.25,
+      "step": 1316
+    },
+    {
+      "epoch": 1.6159509202453988,
+      "grad_norm": 7.935377399082722,
+      "learning_rate": 3.728925619834711e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0349,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.5078125,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.3125,
+      "step": 1317
+    },
+    {
+      "epoch": 1.6171779141104294,
+      "grad_norm": 9.315066544333694,
+      "learning_rate": 3.7256198347107435e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0259,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.431640625,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.0,
+      "step": 1318
+    },
+    {
+      "epoch": 1.61840490797546,
+      "grad_norm": 10.045649594522278,
+      "learning_rate": 3.7223140495867766e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0314,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.8125,
+      "step": 1319
+    },
+    {
+      "epoch": 1.6196319018404908,
+      "grad_norm": 15.320474747736862,
+      "learning_rate": 3.71900826446281e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -378.0,
+      "loss": 0.048,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.3125,
+      "step": 1320
+    },
+    {
+      "epoch": 1.6208588957055214,
+      "grad_norm": 24.6231369547886,
+      "learning_rate": 3.715702479338843e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0705,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.94921875,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.8125,
+      "step": 1321
+    },
+    {
+      "epoch": 1.622085889570552,
+      "grad_norm": 10.301826595324812,
+      "learning_rate": 3.7123966942148756e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0333,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.474609375,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.75,
+      "step": 1322
+    },
+    {
+      "epoch": 1.6233128834355828,
+      "grad_norm": 5.800881401321108,
+      "learning_rate": 3.709090909090909e-07,
+      "logits/chosen": -0.09814453125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -330.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0247,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.66796875,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.6875,
+      "step": 1323
+    },
+    {
+      "epoch": 1.6245398773006134,
+      "grad_norm": 12.535973391030188,
+      "learning_rate": 3.705785123966942e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0588,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.8203125,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.6875,
+      "step": 1324
+    },
+    {
+      "epoch": 1.6257668711656441,
+      "grad_norm": 12.562789704449548,
+      "learning_rate": 3.702479338842975e-07,
+      "logits/chosen": -0.1357421875,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0523,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.353515625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.8125,
+      "step": 1325
+    },
+    {
+      "epoch": 1.6269938650306748,
+      "grad_norm": 7.829908833342415,
+      "learning_rate": 3.6991735537190083e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0269,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.4375,
+      "step": 1326
+    },
+    {
+      "epoch": 1.6282208588957054,
+      "grad_norm": 20.36070349744939,
+      "learning_rate": 3.695867768595041e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0663,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.1201171875,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.6875,
+      "step": 1327
+    },
+    {
+      "epoch": 1.6294478527607361,
+      "grad_norm": 15.025096253851128,
+      "learning_rate": 3.692561983471074e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0478,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.54296875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.25,
+      "step": 1328
+    },
+    {
+      "epoch": 1.6306748466257668,
+      "grad_norm": 11.82653566926932,
+      "learning_rate": 3.6892561983471073e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.30859375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.6875,
+      "step": 1329
+    },
+    {
+      "epoch": 1.6319018404907975,
+      "grad_norm": 10.863416901691561,
+      "learning_rate": 3.6859504132231405e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0477,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.21484375,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -9.0,
+      "step": 1330
+    },
+    {
+      "epoch": 1.6331288343558281,
+      "grad_norm": 10.689179527812403,
+      "learning_rate": 3.682644628099173e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -358.0,
+      "loss": 0.034,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.1455078125,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.875,
+      "step": 1331
+    },
+    {
+      "epoch": 1.6343558282208588,
+      "grad_norm": 10.19931808828623,
+      "learning_rate": 3.6793388429752063e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -364.0,
+      "loss": 0.031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.337890625,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.1875,
+      "step": 1332
+    },
+    {
+      "epoch": 1.6355828220858895,
+      "grad_norm": 16.496319171690203,
+      "learning_rate": 3.6760330578512395e-07,
+      "logits/chosen": -0.341796875,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0631,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.30078125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.0625,
+      "step": 1333
+    },
+    {
+      "epoch": 1.6368098159509201,
+      "grad_norm": 10.236653831572811,
+      "learning_rate": 3.6727272727272727e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0362,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.98828125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.875,
+      "step": 1334
+    },
+    {
+      "epoch": 1.6380368098159508,
+      "grad_norm": 20.7027432046174,
+      "learning_rate": 3.669421487603306e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0518,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.3125,
+      "step": 1335
+    },
+    {
+      "epoch": 1.6392638036809815,
+      "grad_norm": 14.000221716050415,
+      "learning_rate": 3.6661157024793385e-07,
+      "logits/chosen": -0.41796875,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0466,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.064453125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.5,
+      "step": 1336
+    },
+    {
+      "epoch": 1.6404907975460121,
+      "grad_norm": 5.356677692115597,
+      "learning_rate": 3.6628099173553717e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0177,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.0,
+      "step": 1337
+    },
+    {
+      "epoch": 1.6417177914110428,
+      "grad_norm": 9.76155867846443,
+      "learning_rate": 3.659504132231405e-07,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -384.0,
+      "loss": 0.039,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.625,
+      "step": 1338
+    },
+    {
+      "epoch": 1.6429447852760735,
+      "grad_norm": 12.539356287206903,
+      "learning_rate": 3.656198347107438e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -322.0,
+      "loss": 0.0437,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.90625,
+      "step": 1339
+    },
+    {
+      "epoch": 1.6441717791411041,
+      "grad_norm": 18.075978090516806,
+      "learning_rate": 3.6528925619834707e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0864,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 0.83203125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.125,
+      "step": 1340
+    },
+    {
+      "epoch": 1.645398773006135,
+      "grad_norm": 7.930691006707937,
+      "learning_rate": 3.649586776859504e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -434.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0285,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.0625,
+      "step": 1341
+    },
+    {
+      "epoch": 1.6466257668711657,
+      "grad_norm": 7.577446727011239,
+      "learning_rate": 3.646280991735537e-07,
+      "logits/chosen": -0.31640625,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0396,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.92578125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.4375,
+      "step": 1342
+    },
+    {
+      "epoch": 1.6478527607361964,
+      "grad_norm": 33.752974596436516,
+      "learning_rate": 3.64297520661157e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -382.0,
+      "loss": 0.1004,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.373046875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.6875,
+      "step": 1343
+    },
+    {
+      "epoch": 1.649079754601227,
+      "grad_norm": 5.348550645984265,
+      "learning_rate": 3.639669421487603e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0171,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -8.75,
+      "step": 1344
+    },
+    {
+      "epoch": 1.6503067484662577,
+      "grad_norm": 8.28491808447824,
+      "learning_rate": 3.636363636363636e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0326,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7890625,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -7.90625,
+      "step": 1345
+    },
+    {
+      "epoch": 1.6515337423312884,
+      "grad_norm": 18.578868213165027,
+      "learning_rate": 3.633057851239669e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0489,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.4453125,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -8.3125,
+      "step": 1346
+    },
+    {
+      "epoch": 1.652760736196319,
+      "grad_norm": 19.63908403650041,
+      "learning_rate": 3.6297520661157024e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0988,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.4375,
+      "step": 1347
+    },
+    {
+      "epoch": 1.6539877300613497,
+      "grad_norm": 12.572025294012443,
+      "learning_rate": 3.6264462809917356e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -338.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0535,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.2099609375,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.875,
+      "step": 1348
+    },
+    {
+      "epoch": 1.6552147239263804,
+      "grad_norm": 8.973003562840935,
+      "learning_rate": 3.623140495867768e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0377,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.203125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.75,
+      "step": 1349
+    },
+    {
+      "epoch": 1.656441717791411,
+      "grad_norm": 15.006464506327205,
+      "learning_rate": 3.6198347107438014e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -412.0,
+      "loss": 0.05,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.125,
+      "step": 1350
+    },
+    {
+      "epoch": 1.6576687116564417,
+      "grad_norm": 17.74819569728375,
+      "learning_rate": 3.6165289256198346e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0841,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.486328125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.375,
+      "step": 1351
+    },
+    {
+      "epoch": 1.6588957055214724,
+      "grad_norm": 9.171226382603257,
+      "learning_rate": 3.613223140495868e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.470703125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0313,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.4296875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.8125,
+      "step": 1352
+    },
+    {
+      "epoch": 1.660122699386503,
+      "grad_norm": 8.31135298357435,
+      "learning_rate": 3.6099173553719004e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -396.0,
+      "loss": 0.039,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.017578125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -9.125,
+      "step": 1353
+    },
+    {
+      "epoch": 1.6613496932515337,
+      "grad_norm": 9.164109948648726,
+      "learning_rate": 3.6066115702479336e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0391,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.12451171875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.9375,
+      "step": 1354
+    },
+    {
+      "epoch": 1.6625766871165644,
+      "grad_norm": 10.936704965558942,
+      "learning_rate": 3.603305785123967e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0476,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.052978515625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.8125,
+      "step": 1355
+    },
+    {
+      "epoch": 1.6638036809815953,
+      "grad_norm": 13.709468392527326,
+      "learning_rate": 3.6e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.51953125,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -10.1875,
+      "step": 1356
+    },
+    {
+      "epoch": 1.665030674846626,
+      "grad_norm": 9.016472764379131,
+      "learning_rate": 3.596694214876033e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0216,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0673828125,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -10.25,
+      "step": 1357
+    },
+    {
+      "epoch": 1.6662576687116566,
+      "grad_norm": 20.157009863407833,
+      "learning_rate": 3.593388429752066e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0805,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.1181640625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.75,
+      "step": 1358
+    },
+    {
+      "epoch": 1.6674846625766873,
+      "grad_norm": 15.249365615978206,
+      "learning_rate": 3.590082644628099e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0406,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.39453125,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -10.1875,
+      "step": 1359
+    },
+    {
+      "epoch": 1.668711656441718,
+      "grad_norm": 16.208561543286084,
+      "learning_rate": 3.586776859504132e-07,
+      "logits/chosen": -0.45703125,
+      "logits/rejected": -0.69140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0488,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.384765625,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -10.375,
+      "step": 1360
+    },
+    {
+      "epoch": 1.6699386503067486,
+      "grad_norm": 10.803630772069788,
+      "learning_rate": 3.5834710743801653e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0402,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.255859375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.5,
+      "step": 1361
+    },
+    {
+      "epoch": 1.6711656441717793,
+      "grad_norm": 14.627959585498674,
+      "learning_rate": 3.580165289256198e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0524,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.361328125,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.875,
+      "step": 1362
+    },
+    {
+      "epoch": 1.67239263803681,
+      "grad_norm": 4.581091718122957,
+      "learning_rate": 3.576859504132231e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0148,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0791015625,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -10.5,
+      "step": 1363
+    },
+    {
+      "epoch": 1.6736196319018406,
+      "grad_norm": 8.359592521972438,
+      "learning_rate": 3.5735537190082643e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0237,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.515625,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.75,
+      "step": 1364
+    },
+    {
+      "epoch": 1.6748466257668713,
+      "grad_norm": 16.525640852041033,
+      "learning_rate": 3.5702479338842975e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0676,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.65625,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.5625,
+      "step": 1365
+    },
+    {
+      "epoch": 1.676073619631902,
+      "grad_norm": 7.921394403782825,
+      "learning_rate": 3.5669421487603307e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0262,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.83984375,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.3125,
+      "step": 1366
+    },
+    {
+      "epoch": 1.6773006134969326,
+      "grad_norm": 12.021872239150174,
+      "learning_rate": 3.5636363636363633e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.7734375,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.8125,
+      "step": 1367
+    },
+    {
+      "epoch": 1.6785276073619633,
+      "grad_norm": 18.310359770819986,
+      "learning_rate": 3.5603305785123965e-07,
+      "logits/chosen": -0.376953125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -378.0,
+      "loss": 0.057,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1357421875,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.9375,
+      "step": 1368
+    },
+    {
+      "epoch": 1.679754601226994,
+      "grad_norm": 16.65697818828575,
+      "learning_rate": 3.5570247933884297e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -380.0,
+      "loss": 0.079,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.0218505859375,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.3125,
+      "step": 1369
+    },
+    {
+      "epoch": 1.6809815950920246,
+      "grad_norm": 11.204270470130796,
+      "learning_rate": 3.553719008264463e-07,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0306,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.75,
+      "step": 1370
+    },
+    {
+      "epoch": 1.6822085889570553,
+      "grad_norm": 15.13195763037223,
+      "learning_rate": 3.5504132231404955e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -464.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0514,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -9.5,
+      "step": 1371
+    },
+    {
+      "epoch": 1.683435582822086,
+      "grad_norm": 8.16317454264987,
+      "learning_rate": 3.5471074380165287e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0346,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.4375,
+      "step": 1372
+    },
+    {
+      "epoch": 1.6846625766871166,
+      "grad_norm": 15.473022951774299,
+      "learning_rate": 3.543801652892562e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -394.0,
+      "loss": 0.05,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7265625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.0625,
+      "step": 1373
+    },
+    {
+      "epoch": 1.6858895705521473,
+      "grad_norm": 13.313790214938209,
+      "learning_rate": 3.540495867768595e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0592,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.3125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.6875,
+      "step": 1374
+    },
+    {
+      "epoch": 1.687116564417178,
+      "grad_norm": 15.234502433567197,
+      "learning_rate": 3.537190082644628e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -326.0,
+      "logps/rejected": -342.0,
+      "loss": 0.065,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.498046875,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.625,
+      "step": 1375
+    },
+    {
+      "epoch": 1.6883435582822086,
+      "grad_norm": 7.451858103319708,
+      "learning_rate": 3.533884297520661e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0294,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.375,
+      "step": 1376
+    },
+    {
+      "epoch": 1.6895705521472393,
+      "grad_norm": 11.164973482876846,
+      "learning_rate": 3.530578512396694e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0356,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -9.5625,
+      "step": 1377
+    },
+    {
+      "epoch": 1.69079754601227,
+      "grad_norm": 7.087310188668901,
+      "learning_rate": 3.527272727272727e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0308,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.193359375,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -8.5625,
+      "step": 1378
+    },
+    {
+      "epoch": 1.6920245398773006,
+      "grad_norm": 9.503069803683369,
+      "learning_rate": 3.5239669421487604e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0483,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -7.96875,
+      "step": 1379
+    },
+    {
+      "epoch": 1.6932515337423313,
+      "grad_norm": 6.484416144875946,
+      "learning_rate": 3.520661157024793e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0227,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.3125,
+      "step": 1380
+    },
+    {
+      "epoch": 1.694478527607362,
+      "grad_norm": 12.461259591914391,
+      "learning_rate": 3.517355371900826e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.1767578125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0506,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.482421875,
+      "rewards/margins": 8.4375,
+      "rewards/rejected": -7.9375,
+      "step": 1381
+    },
+    {
+      "epoch": 1.6957055214723926,
+      "grad_norm": 12.727643966429453,
+      "learning_rate": 3.5140495867768594e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0412,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2294921875,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.625,
+      "step": 1382
+    },
+    {
+      "epoch": 1.6969325153374233,
+      "grad_norm": 6.275694798930385,
+      "learning_rate": 3.5107438016528925e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0245,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.5625,
+      "step": 1383
+    },
+    {
+      "epoch": 1.698159509202454,
+      "grad_norm": 13.180464183457763,
+      "learning_rate": 3.5074380165289257e-07,
+      "logits/chosen": -0.095703125,
+      "logits/rejected": -0.21484375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0482,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.51171875,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.0625,
+      "step": 1384
+    },
+    {
+      "epoch": 1.6993865030674846,
+      "grad_norm": 13.726588628334937,
+      "learning_rate": 3.5041322314049584e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -326.0,
+      "logps/rejected": -326.0,
+      "loss": 0.058,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.68359375,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.25,
+      "step": 1385
+    },
+    {
+      "epoch": 1.7006134969325153,
+      "grad_norm": 9.154420481281178,
+      "learning_rate": 3.5008264462809915e-07,
+      "logits/chosen": -0.359375,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0368,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.0,
+      "step": 1386
+    },
+    {
+      "epoch": 1.701840490797546,
+      "grad_norm": 11.579193730406892,
+      "learning_rate": 3.4975206611570247e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0428,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.0,
+      "step": 1387
+    },
+    {
+      "epoch": 1.7030674846625766,
+      "grad_norm": 9.159728350094172,
+      "learning_rate": 3.494214876033058e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0348,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.390625,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -8.1875,
+      "step": 1388
+    },
+    {
+      "epoch": 1.7042944785276073,
+      "grad_norm": 10.976035161259295,
+      "learning_rate": 3.4909090909090905e-07,
+      "logits/chosen": -0.361328125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0444,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9140625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.9375,
+      "step": 1389
+    },
+    {
+      "epoch": 1.705521472392638,
+      "grad_norm": 15.764967882572376,
+      "learning_rate": 3.4876033057851237e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0452,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.58984375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.125,
+      "step": 1390
+    },
+    {
+      "epoch": 1.7067484662576686,
+      "grad_norm": 7.403388918790972,
+      "learning_rate": 3.484297520661157e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0356,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.375,
+      "step": 1391
+    },
+    {
+      "epoch": 1.7079754601226993,
+      "grad_norm": 10.896245766912692,
+      "learning_rate": 3.48099173553719e-07,
+      "logits/chosen": -0.10107421875,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.31640625,
+      "rewards/margins": 8.1875,
+      "rewards/rejected": -7.875,
+      "step": 1392
+    },
+    {
+      "epoch": 1.70920245398773,
+      "grad_norm": 11.366422205322182,
+      "learning_rate": 3.4776859504132227e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -328.0,
+      "loss": 0.0456,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.0625,
+      "step": 1393
+    },
+    {
+      "epoch": 1.7104294478527606,
+      "grad_norm": 14.99148246159462,
+      "learning_rate": 3.474380165289256e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0628,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.8125,
+      "step": 1394
+    },
+    {
+      "epoch": 1.7116564417177913,
+      "grad_norm": 5.724911703113608,
+      "learning_rate": 3.471074380165289e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0234,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.25,
+      "step": 1395
+    },
+    {
+      "epoch": 1.712883435582822,
+      "grad_norm": 7.299931111351364,
+      "learning_rate": 3.467768595041322e-07,
+      "logits/chosen": -0.0869140625,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0281,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.953125,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.125,
+      "step": 1396
+    },
+    {
+      "epoch": 1.7141104294478526,
+      "grad_norm": 7.734982226412597,
+      "learning_rate": 3.4644628099173554e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0299,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.1875,
+      "step": 1397
+    },
+    {
+      "epoch": 1.7153374233128833,
+      "grad_norm": 12.433607504676552,
+      "learning_rate": 3.461157024793388e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0423,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.6875,
+      "step": 1398
+    },
+    {
+      "epoch": 1.716564417177914,
+      "grad_norm": 10.213873529131746,
+      "learning_rate": 3.457851239669421e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0372,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.494140625,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -8.625,
+      "step": 1399
+    },
+    {
+      "epoch": 1.7177914110429446,
+      "grad_norm": 8.219886818141926,
+      "learning_rate": 3.4545454545454544e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -340.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0241,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.025146484375,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.8125,
+      "step": 1400
+    },
+    {
+      "epoch": 1.7190184049079753,
+      "grad_norm": 8.611806636266707,
+      "learning_rate": 3.4512396694214876e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -444.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0354,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.25,
+      "step": 1401
+    },
+    {
+      "epoch": 1.720245398773006,
+      "grad_norm": 9.897114375609455,
+      "learning_rate": 3.44793388429752e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0394,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.578125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.9375,
+      "step": 1402
+    },
+    {
+      "epoch": 1.7214723926380369,
+      "grad_norm": 8.741002963663481,
+      "learning_rate": 3.4446280991735534e-07,
+      "logits/chosen": -0.1572265625,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0278,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.46484375,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.8125,
+      "step": 1403
+    },
+    {
+      "epoch": 1.7226993865030675,
+      "grad_norm": 10.367535092764449,
+      "learning_rate": 3.4413223140495866e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -334.0,
+      "logps/rejected": -308.0,
+      "loss": 0.0367,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.91796875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.6875,
+      "step": 1404
+    },
+    {
+      "epoch": 1.7239263803680982,
+      "grad_norm": 13.241758246923336,
+      "learning_rate": 3.43801652892562e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0467,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.55859375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.375,
+      "step": 1405
+    },
+    {
+      "epoch": 1.7251533742331289,
+      "grad_norm": 8.331965679319266,
+      "learning_rate": 3.434710743801653e-07,
+      "logits/chosen": -0.38671875,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -324.0,
+      "logps/rejected": -352.0,
+      "loss": 0.023,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.01019287109375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.9375,
+      "step": 1406
+    },
+    {
+      "epoch": 1.7263803680981595,
+      "grad_norm": 15.380429512755986,
+      "learning_rate": 3.4314049586776856e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -438.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0578,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -8.9375,
+      "step": 1407
+    },
+    {
+      "epoch": 1.7276073619631902,
+      "grad_norm": 11.962024064286275,
+      "learning_rate": 3.428099173553719e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -470.0,
+      "logps/rejected": -410.0,
+      "loss": 0.058,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.921875,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.1875,
+      "step": 1408
+    },
+    {
+      "epoch": 1.7288343558282209,
+      "grad_norm": 12.743570008480159,
+      "learning_rate": 3.424793388429752e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.330078125,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.25,
+      "step": 1409
+    },
+    {
+      "epoch": 1.7300613496932515,
+      "grad_norm": 10.912975195599188,
+      "learning_rate": 3.421487603305785e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0529,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.05126953125,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -9.0,
+      "step": 1410
+    },
+    {
+      "epoch": 1.7312883435582822,
+      "grad_norm": 9.273924247986834,
+      "learning_rate": 3.418181818181818e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0371,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.0213623046875,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.375,
+      "step": 1411
+    },
+    {
+      "epoch": 1.7325153374233129,
+      "grad_norm": 16.100998085595243,
+      "learning_rate": 3.414876033057851e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0693,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.66015625,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.1875,
+      "step": 1412
+    },
+    {
+      "epoch": 1.7337423312883435,
+      "grad_norm": 12.422653573930514,
+      "learning_rate": 3.411570247933884e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0534,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.037353515625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.6875,
+      "step": 1413
+    },
+    {
+      "epoch": 1.7349693251533742,
+      "grad_norm": 8.035094608949246,
+      "learning_rate": 3.4082644628099173e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -432.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0358,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.5546875,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.0,
+      "step": 1414
+    },
+    {
+      "epoch": 1.7361963190184049,
+      "grad_norm": 15.842862209397916,
+      "learning_rate": 3.4049586776859505e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0571,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.265625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.25,
+      "step": 1415
+    },
+    {
+      "epoch": 1.7374233128834355,
+      "grad_norm": 12.823306849806624,
+      "learning_rate": 3.401652892561983e-07,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0456,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9921875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.0625,
+      "step": 1416
+    },
+    {
+      "epoch": 1.7386503067484662,
+      "grad_norm": 9.156727463682909,
+      "learning_rate": 3.3983471074380163e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0337,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5546875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.5,
+      "step": 1417
+    },
+    {
+      "epoch": 1.7398773006134969,
+      "grad_norm": 11.45003765016041,
+      "learning_rate": 3.3950413223140495e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0371,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.3125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.5625,
+      "step": 1418
+    },
+    {
+      "epoch": 1.7411042944785278,
+      "grad_norm": 12.920796689779438,
+      "learning_rate": 3.3917355371900827e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -340.0,
+      "logps/rejected": -318.0,
+      "loss": 0.0579,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.11572265625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.75,
+      "step": 1419
+    },
+    {
+      "epoch": 1.7423312883435584,
+      "grad_norm": 14.910027932916357,
+      "learning_rate": 3.3884297520661153e-07,
+      "logits/chosen": -0.396484375,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -396.0,
+      "loss": 0.039,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.57421875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.8125,
+      "step": 1420
+    },
+    {
+      "epoch": 1.743558282208589,
+      "grad_norm": 8.893525346196068,
+      "learning_rate": 3.3851239669421485e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.5,
+      "logps/chosen": -376.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0345,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.53515625,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.875,
+      "step": 1421
+    },
+    {
+      "epoch": 1.7447852760736198,
+      "grad_norm": 15.365028675388714,
+      "learning_rate": 3.3818181818181817e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0651,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.3046875,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.9375,
+      "step": 1422
+    },
+    {
+      "epoch": 1.7460122699386504,
+      "grad_norm": 9.21774061389196,
+      "learning_rate": 3.378512396694215e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0397,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.3828125,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -10.5625,
+      "step": 1423
+    },
+    {
+      "epoch": 1.747239263803681,
+      "grad_norm": 10.515642559799373,
+      "learning_rate": 3.375206611570248e-07,
+      "logits/chosen": -0.330078125,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0431,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.333984375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.625,
+      "step": 1424
+    },
+    {
+      "epoch": 1.7484662576687118,
+      "grad_norm": 10.498215748420622,
+      "learning_rate": 3.3719008264462807e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0385,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.431640625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.625,
+      "step": 1425
+    },
+    {
+      "epoch": 1.7496932515337424,
+      "grad_norm": 13.42107772359636,
+      "learning_rate": 3.368595041322314e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0665,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.427734375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.25,
+      "step": 1426
+    },
+    {
+      "epoch": 1.7509202453987731,
+      "grad_norm": 17.515697940964518,
+      "learning_rate": 3.365289256198347e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0599,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -8.5625,
+      "step": 1427
+    },
+    {
+      "epoch": 1.7521472392638038,
+      "grad_norm": 12.347321878732972,
+      "learning_rate": 3.36198347107438e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0456,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9296875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.3125,
+      "step": 1428
+    },
+    {
+      "epoch": 1.7533742331288344,
+      "grad_norm": 11.083184332644388,
+      "learning_rate": 3.358677685950413e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0441,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.59375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.0,
+      "step": 1429
+    },
+    {
+      "epoch": 1.7546012269938651,
+      "grad_norm": 7.888788470102411,
+      "learning_rate": 3.355371900826446e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0322,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.58203125,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.1875,
+      "step": 1430
+    },
+    {
+      "epoch": 1.7558282208588958,
+      "grad_norm": 10.25363004538014,
+      "learning_rate": 3.352066115702479e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0383,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.44921875,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.0625,
+      "step": 1431
+    },
+    {
+      "epoch": 1.7570552147239265,
+      "grad_norm": 5.057831122485827,
+      "learning_rate": 3.3487603305785124e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -334.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0153,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.25,
+      "step": 1432
+    },
+    {
+      "epoch": 1.7582822085889571,
+      "grad_norm": 17.51301080467881,
+      "learning_rate": 3.345454545454545e-07,
+      "logits/chosen": -0.10107421875,
+      "logits/rejected": -0.21484375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0887,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.13671875,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -8.0,
+      "step": 1433
+    },
+    {
+      "epoch": 1.7595092024539878,
+      "grad_norm": 11.377580450972712,
+      "learning_rate": 3.342148760330578e-07,
+      "logits/chosen": -0.126953125,
+      "logits/rejected": -0.2275390625,
+      "logps/chosen": -342.0,
+      "logps/rejected": -390.0,
+      "loss": 0.045,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.357421875,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.25,
+      "step": 1434
+    },
+    {
+      "epoch": 1.7607361963190185,
+      "grad_norm": 8.026386246513397,
+      "learning_rate": 3.3388429752066114e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -348.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0306,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.57421875,
+      "rewards/margins": 8.9375,
+      "rewards/rejected": -8.375,
+      "step": 1435
+    },
+    {
+      "epoch": 1.7619631901840491,
+      "grad_norm": 5.610915011446518,
+      "learning_rate": 3.3355371900826446e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0205,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -8.875,
+      "step": 1436
+    },
+    {
+      "epoch": 1.7631901840490798,
+      "grad_norm": 12.950887212243945,
+      "learning_rate": 3.332231404958678e-07,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0445,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.1875,
+      "step": 1437
+    },
+    {
+      "epoch": 1.7644171779141105,
+      "grad_norm": 12.771531747113794,
+      "learning_rate": 3.3289256198347104e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0399,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.42578125,
+      "rewards/margins": 8.875,
+      "rewards/rejected": -8.4375,
+      "step": 1438
+    },
+    {
+      "epoch": 1.7656441717791411,
+      "grad_norm": 9.19835013594589,
+      "learning_rate": 3.3256198347107436e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0317,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.027099609375,
+      "rewards/margins": 8.5625,
+      "rewards/rejected": -8.5625,
+      "step": 1439
+    },
+    {
+      "epoch": 1.7668711656441718,
+      "grad_norm": 7.045378068044565,
+      "learning_rate": 3.322314049586777e-07,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -466.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0232,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.25,
+      "step": 1440
+    },
+    {
+      "epoch": 1.7680981595092025,
+      "grad_norm": 11.45812069017029,
+      "learning_rate": 3.31900826446281e-07,
+      "logits/chosen": -0.37890625,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0426,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.443359375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.625,
+      "step": 1441
+    },
+    {
+      "epoch": 1.7693251533742331,
+      "grad_norm": 15.329344086941836,
+      "learning_rate": 3.3157024793388426e-07,
+      "logits/chosen": -0.322265625,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -430.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0525,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.490234375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.125,
+      "step": 1442
+    },
+    {
+      "epoch": 1.7705521472392638,
+      "grad_norm": 8.1454122193256,
+      "learning_rate": 3.312396694214876e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0342,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9375,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.1875,
+      "step": 1443
+    },
+    {
+      "epoch": 1.7717791411042945,
+      "grad_norm": 18.4128990711308,
+      "learning_rate": 3.309090909090909e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0418,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.0625,
+      "step": 1444
+    },
+    {
+      "epoch": 1.7730061349693251,
+      "grad_norm": 10.577537024132393,
+      "learning_rate": 3.305785123966942e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -326.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0331,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1025390625,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -9.0625,
+      "step": 1445
+    },
+    {
+      "epoch": 1.7742331288343558,
+      "grad_norm": 14.48140279792318,
+      "learning_rate": 3.3024793388429753e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0426,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.486328125,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -10.0,
+      "step": 1446
+    },
+    {
+      "epoch": 1.7754601226993865,
+      "grad_norm": 10.710617346705183,
+      "learning_rate": 3.299173553719008e-07,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0422,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1640625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.5625,
+      "step": 1447
+    },
+    {
+      "epoch": 1.7766871165644171,
+      "grad_norm": 13.124515407591634,
+      "learning_rate": 3.295867768595041e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -342.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0472,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.3359375,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -9.375,
+      "step": 1448
+    },
+    {
+      "epoch": 1.7779141104294478,
+      "grad_norm": 11.195883118200678,
+      "learning_rate": 3.2925619834710743e-07,
+      "logits/chosen": -0.10595703125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0415,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.2490234375,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -9.0,
+      "step": 1449
+    },
+    {
+      "epoch": 1.7791411042944785,
+      "grad_norm": 13.381809497018025,
+      "learning_rate": 3.2892561983471075e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.421875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.3125,
+      "step": 1450
+    },
+    {
+      "epoch": 1.7803680981595091,
+      "grad_norm": 9.53349976877088,
+      "learning_rate": 3.28595041322314e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0383,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.5390625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.3125,
+      "step": 1451
+    },
+    {
+      "epoch": 1.7815950920245398,
+      "grad_norm": 14.85880299692277,
+      "learning_rate": 3.2826446280991733e-07,
+      "logits/chosen": -0.373046875,
+      "logits/rejected": -0.625,
+      "logps/chosen": -454.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0233,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -9.75,
+      "step": 1452
+    },
+    {
+      "epoch": 1.7828220858895705,
+      "grad_norm": 7.4366621556306605,
+      "learning_rate": 3.2793388429752065e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0244,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.8125,
+      "step": 1453
+    },
+    {
+      "epoch": 1.7840490797546011,
+      "grad_norm": 14.1195355778993,
+      "learning_rate": 3.2760330578512396e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0325,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.27734375,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.9375,
+      "step": 1454
+    },
+    {
+      "epoch": 1.7852760736196318,
+      "grad_norm": 9.184961337980399,
+      "learning_rate": 3.272727272727273e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0351,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.875,
+      "step": 1455
+    },
+    {
+      "epoch": 1.7865030674846625,
+      "grad_norm": 11.51120353006608,
+      "learning_rate": 3.2694214876033055e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0465,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.25390625,
+      "rewards/margins": 8.625,
+      "rewards/rejected": -8.375,
+      "step": 1456
+    },
+    {
+      "epoch": 1.7877300613496931,
+      "grad_norm": 10.80392052881998,
+      "learning_rate": 3.2661157024793386e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -420.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0384,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.455078125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.4375,
+      "step": 1457
+    },
+    {
+      "epoch": 1.7889570552147238,
+      "grad_norm": 19.370928970130752,
+      "learning_rate": 3.262809917355372e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.2421875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0589,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.04345703125,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -8.4375,
+      "step": 1458
+    },
+    {
+      "epoch": 1.7901840490797545,
+      "grad_norm": 8.14401503004669,
+      "learning_rate": 3.259504132231405e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0249,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.73828125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.6875,
+      "step": 1459
+    },
+    {
+      "epoch": 1.7914110429447851,
+      "grad_norm": 8.410071233667539,
+      "learning_rate": 3.2561983471074376e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0337,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.796875,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -8.8125,
+      "step": 1460
+    },
+    {
+      "epoch": 1.7926380368098158,
+      "grad_norm": 17.319232365393777,
+      "learning_rate": 3.252892561983471e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0584,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8203125,
+      "rewards/margins": 11.125,
+      "rewards/rejected": -10.3125,
+      "step": 1461
+    },
+    {
+      "epoch": 1.7938650306748465,
+      "grad_norm": 11.658403751664542,
+      "learning_rate": 3.249586776859504e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -324.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0423,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.412109375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.125,
+      "step": 1462
+    },
+    {
+      "epoch": 1.7950920245398772,
+      "grad_norm": 6.975630646192431,
+      "learning_rate": 3.246280991735537e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0238,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.65625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -8.875,
+      "step": 1463
+    },
+    {
+      "epoch": 1.7963190184049078,
+      "grad_norm": 13.732283507159519,
+      "learning_rate": 3.2429752066115704e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0461,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.294921875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.5,
+      "step": 1464
+    },
+    {
+      "epoch": 1.7975460122699385,
+      "grad_norm": 9.668785840082787,
+      "learning_rate": 3.239669421487603e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0301,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.6171875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.625,
+      "step": 1465
+    },
+    {
+      "epoch": 1.7987730061349694,
+      "grad_norm": 5.445268106615218,
+      "learning_rate": 3.236363636363636e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0153,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1328125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.8125,
+      "step": 1466
+    },
+    {
+      "epoch": 1.8,
+      "grad_norm": 8.337653458750603,
+      "learning_rate": 3.2330578512396694e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0287,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -8.625,
+      "step": 1467
+    },
+    {
+      "epoch": 1.8012269938650307,
+      "grad_norm": 14.119670175365039,
+      "learning_rate": 3.2297520661157025e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0381,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.0625,
+      "step": 1468
+    },
+    {
+      "epoch": 1.8024539877300614,
+      "grad_norm": 15.57267164353982,
+      "learning_rate": 3.226446280991735e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0642,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.32421875,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.375,
+      "step": 1469
+    },
+    {
+      "epoch": 1.803680981595092,
+      "grad_norm": 13.78094405933175,
+      "learning_rate": 3.2231404958677684e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0781,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.022216796875,
+      "rewards/margins": 8.8125,
+      "rewards/rejected": -8.8125,
+      "step": 1470
+    },
+    {
+      "epoch": 1.8049079754601227,
+      "grad_norm": 8.008772562359416,
+      "learning_rate": 3.2198347107438015e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -374.0,
+      "loss": 0.03,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -8.4375,
+      "step": 1471
+    },
+    {
+      "epoch": 1.8061349693251534,
+      "grad_norm": 15.98977849235284,
+      "learning_rate": 3.2165289256198347e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0514,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.75,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.125,
+      "step": 1472
+    },
+    {
+      "epoch": 1.807361963190184,
+      "grad_norm": 9.291346547833697,
+      "learning_rate": 3.2132231404958674e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -362.0,
+      "loss": 0.037,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.83984375,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.5,
+      "step": 1473
+    },
+    {
+      "epoch": 1.8085889570552147,
+      "grad_norm": 18.454994761707635,
+      "learning_rate": 3.2099173553719005e-07,
+      "logits/chosen": -0.10107421875,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0443,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.9609375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.0,
+      "step": 1474
+    },
+    {
+      "epoch": 1.8098159509202454,
+      "grad_norm": 9.644230934604456,
+      "learning_rate": 3.2066115702479337e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0343,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7265625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.0,
+      "step": 1475
+    },
+    {
+      "epoch": 1.811042944785276,
+      "grad_norm": 6.913662419393237,
+      "learning_rate": 3.203305785123967e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0234,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.3125,
+      "step": 1476
+    },
+    {
+      "epoch": 1.8122699386503067,
+      "grad_norm": 13.89342979775928,
+      "learning_rate": 3.2e-07,
+      "logits/chosen": -0.1396484375,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0585,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.419921875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.0,
+      "step": 1477
+    },
+    {
+      "epoch": 1.8134969325153374,
+      "grad_norm": 12.2206513449555,
+      "learning_rate": 3.1966942148760327e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0407,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.4296875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -8.8125,
+      "step": 1478
+    },
+    {
+      "epoch": 1.814723926380368,
+      "grad_norm": 17.54721691614464,
+      "learning_rate": 3.193388429752066e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0468,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.88671875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.5,
+      "step": 1479
+    },
+    {
+      "epoch": 1.8159509202453987,
+      "grad_norm": 18.143441483594128,
+      "learning_rate": 3.190082644628099e-07,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -392.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0632,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.25,
+      "step": 1480
+    },
+    {
+      "epoch": 1.8171779141104294,
+      "grad_norm": 16.606911484964844,
+      "learning_rate": 3.186776859504132e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.384765625,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.5625,
+      "step": 1481
+    },
+    {
+      "epoch": 1.8184049079754603,
+      "grad_norm": 13.154293841882401,
+      "learning_rate": 3.183471074380165e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0766,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.5390625,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -8.6875,
+      "step": 1482
+    },
+    {
+      "epoch": 1.819631901840491,
+      "grad_norm": 14.861363357929438,
+      "learning_rate": 3.180165289256198e-07,
+      "logits/chosen": -0.0830078125,
+      "logits/rejected": -0.2431640625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -316.0,
+      "loss": 0.0629,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.345703125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.6875,
+      "step": 1483
+    },
+    {
+      "epoch": 1.8208588957055216,
+      "grad_norm": 14.681356413204243,
+      "learning_rate": 3.176859504132231e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -348.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0461,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.1376953125,
+      "rewards/margins": 9.125,
+      "rewards/rejected": -9.0,
+      "step": 1484
+    },
+    {
+      "epoch": 1.8220858895705523,
+      "grad_norm": 10.743142311605123,
+      "learning_rate": 3.1735537190082644e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0294,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.004547119140625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.3125,
+      "step": 1485
+    },
+    {
+      "epoch": 1.823312883435583,
+      "grad_norm": 15.212920486790656,
+      "learning_rate": 3.1702479338842976e-07,
+      "logits/chosen": -0.353515625,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0613,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.412109375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.625,
+      "step": 1486
+    },
+    {
+      "epoch": 1.8245398773006136,
+      "grad_norm": 7.09810262598987,
+      "learning_rate": 3.16694214876033e-07,
+      "logits/chosen": -0.40234375,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0222,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.875,
+      "step": 1487
+    },
+    {
+      "epoch": 1.8257668711656443,
+      "grad_norm": 13.16618442532596,
+      "learning_rate": 3.1636363636363634e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0301,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.734375,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.6875,
+      "step": 1488
+    },
+    {
+      "epoch": 1.826993865030675,
+      "grad_norm": 13.779672103672336,
+      "learning_rate": 3.1603305785123966e-07,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -380.0,
+      "loss": 0.051,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.796875,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -9.875,
+      "step": 1489
+    },
+    {
+      "epoch": 1.8282208588957056,
+      "grad_norm": 11.957973151160742,
+      "learning_rate": 3.15702479338843e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0395,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.84375,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.375,
+      "step": 1490
+    },
+    {
+      "epoch": 1.8294478527607363,
+      "grad_norm": 12.913847461598793,
+      "learning_rate": 3.1537190082644624e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -480.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0464,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -9.6875,
+      "step": 1491
+    },
+    {
+      "epoch": 1.830674846625767,
+      "grad_norm": 13.516942361236422,
+      "learning_rate": 3.1504132231404956e-07,
+      "logits/chosen": -0.412109375,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0473,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.126953125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.1875,
+      "step": 1492
+    },
+    {
+      "epoch": 1.8319018404907976,
+      "grad_norm": 9.02996772401457,
+      "learning_rate": 3.147107438016529e-07,
+      "logits/chosen": -0.365234375,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0384,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.357421875,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -9.5625,
+      "step": 1493
+    },
+    {
+      "epoch": 1.8331288343558283,
+      "grad_norm": 7.984291330585717,
+      "learning_rate": 3.143801652892562e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -438.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0342,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.318359375,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.6875,
+      "step": 1494
+    },
+    {
+      "epoch": 1.834355828220859,
+      "grad_norm": 12.136279210243362,
+      "learning_rate": 3.140495867768595e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0394,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.462890625,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -8.875,
+      "step": 1495
+    },
+    {
+      "epoch": 1.8355828220858896,
+      "grad_norm": 7.676908976526212,
+      "learning_rate": 3.137190082644628e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0219,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.80859375,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.0625,
+      "step": 1496
+    },
+    {
+      "epoch": 1.8368098159509203,
+      "grad_norm": 6.418224453171171,
+      "learning_rate": 3.133884297520661e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0244,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -8.9375,
+      "step": 1497
+    },
+    {
+      "epoch": 1.838036809815951,
+      "grad_norm": 14.435468659209993,
+      "learning_rate": 3.130578512396694e-07,
+      "logits/chosen": -0.37890625,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0507,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.392578125,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.0625,
+      "step": 1498
+    },
+    {
+      "epoch": 1.8392638036809816,
+      "grad_norm": 7.358594644326529,
+      "learning_rate": 3.1272727272727273e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0276,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.89453125,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.0,
+      "step": 1499
+    },
+    {
+      "epoch": 1.8404907975460123,
+      "grad_norm": 12.861833790262308,
+      "learning_rate": 3.12396694214876e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0584,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -8.625,
+      "step": 1500
+    },
+    {
+      "epoch": 1.841717791411043,
+      "grad_norm": 9.440131410139509,
+      "learning_rate": 3.120661157024793e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -376.0,
+      "loss": 0.036,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.0,
+      "step": 1501
+    },
+    {
+      "epoch": 1.8429447852760736,
+      "grad_norm": 10.759450623282584,
+      "learning_rate": 3.1173553719008263e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0313,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.24609375,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -9.4375,
+      "step": 1502
+    },
+    {
+      "epoch": 1.8441717791411043,
+      "grad_norm": 13.44299298290354,
+      "learning_rate": 3.1140495867768595e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0429,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.375,
+      "step": 1503
+    },
+    {
+      "epoch": 1.845398773006135,
+      "grad_norm": 10.171235600186767,
+      "learning_rate": 3.1107438016528927e-07,
+      "logits/chosen": -0.322265625,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0353,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.125,
+      "step": 1504
+    },
+    {
+      "epoch": 1.8466257668711656,
+      "grad_norm": 10.219435819353786,
+      "learning_rate": 3.1074380165289253e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -398.0,
+      "loss": 0.04,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.220703125,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.125,
+      "step": 1505
+    },
+    {
+      "epoch": 1.8478527607361963,
+      "grad_norm": 8.678347100798598,
+      "learning_rate": 3.1041322314049585e-07,
+      "logits/chosen": -0.248046875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0319,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.0145263671875,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.3125,
+      "step": 1506
+    },
+    {
+      "epoch": 1.849079754601227,
+      "grad_norm": 12.945184798310114,
+      "learning_rate": 3.1008264462809917e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0488,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.87890625,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.375,
+      "step": 1507
+    },
+    {
+      "epoch": 1.8503067484662576,
+      "grad_norm": 10.145841916730342,
+      "learning_rate": 3.097520661157025e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0347,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1806640625,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.125,
+      "step": 1508
+    },
+    {
+      "epoch": 1.8515337423312883,
+      "grad_norm": 9.354504262490346,
+      "learning_rate": 3.0942148760330575e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0279,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -8.8125,
+      "step": 1509
+    },
+    {
+      "epoch": 1.852760736196319,
+      "grad_norm": 9.721886786793712,
+      "learning_rate": 3.0909090909090907e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0299,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.75,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.4375,
+      "step": 1510
+    },
+    {
+      "epoch": 1.8539877300613496,
+      "grad_norm": 6.801235608742616,
+      "learning_rate": 3.087603305785124e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0197,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -10.0625,
+      "step": 1511
+    },
+    {
+      "epoch": 1.8552147239263803,
+      "grad_norm": 9.356069818038527,
+      "learning_rate": 3.084297520661157e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0352,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -9.6875,
+      "step": 1512
+    },
+    {
+      "epoch": 1.856441717791411,
+      "grad_norm": 15.1733760873437,
+      "learning_rate": 3.08099173553719e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0447,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.337890625,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.375,
+      "step": 1513
+    },
+    {
+      "epoch": 1.8576687116564417,
+      "grad_norm": 8.210480914006832,
+      "learning_rate": 3.077685950413223e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0348,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.671875,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -10.5625,
+      "step": 1514
+    },
+    {
+      "epoch": 1.8588957055214723,
+      "grad_norm": 15.498698060895602,
+      "learning_rate": 3.074380165289256e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0445,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -10.0,
+      "step": 1515
+    },
+    {
+      "epoch": 1.860122699386503,
+      "grad_norm": 11.441149191311155,
+      "learning_rate": 3.071074380165289e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0493,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.3671875,
+      "rewards/margins": 9.1875,
+      "rewards/rejected": -9.5625,
+      "step": 1516
+    },
+    {
+      "epoch": 1.8613496932515337,
+      "grad_norm": 6.593493238446448,
+      "learning_rate": 3.0677685950413224e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0262,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.84765625,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -9.9375,
+      "step": 1517
+    },
+    {
+      "epoch": 1.8625766871165643,
+      "grad_norm": 13.313579919939231,
+      "learning_rate": 3.064462809917355e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0479,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.21875,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -10.25,
+      "step": 1518
+    },
+    {
+      "epoch": 1.863803680981595,
+      "grad_norm": 18.06842961008156,
+      "learning_rate": 3.061157024793388e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0436,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.67578125,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -9.9375,
+      "step": 1519
+    },
+    {
+      "epoch": 1.8650306748466257,
+      "grad_norm": 6.182876979289249,
+      "learning_rate": 3.0578512396694214e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0203,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.026611328125,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -10.25,
+      "step": 1520
+    },
+    {
+      "epoch": 1.8662576687116563,
+      "grad_norm": 13.565119222091443,
+      "learning_rate": 3.0545454545454546e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.046630859375,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.375,
+      "step": 1521
+    },
+    {
+      "epoch": 1.867484662576687,
+      "grad_norm": 11.55315965721378,
+      "learning_rate": 3.051239669421487e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0499,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.546875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.5,
+      "step": 1522
+    },
+    {
+      "epoch": 1.8687116564417177,
+      "grad_norm": 8.768625848422289,
+      "learning_rate": 3.0479338842975204e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0266,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.515625,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.75,
+      "step": 1523
+    },
+    {
+      "epoch": 1.8699386503067483,
+      "grad_norm": 17.46967243494328,
+      "learning_rate": 3.0446280991735536e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0648,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.47265625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.9375,
+      "step": 1524
+    },
+    {
+      "epoch": 1.871165644171779,
+      "grad_norm": 18.901627284827356,
+      "learning_rate": 3.041322314049587e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0553,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -10.0,
+      "step": 1525
+    },
+    {
+      "epoch": 1.8723926380368097,
+      "grad_norm": 15.46265759049967,
+      "learning_rate": 3.03801652892562e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0669,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -8.9375,
+      "step": 1526
+    },
+    {
+      "epoch": 1.8736196319018403,
+      "grad_norm": 15.793631466248941,
+      "learning_rate": 3.0347107438016526e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0551,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.125,
+      "step": 1527
+    },
+    {
+      "epoch": 1.874846625766871,
+      "grad_norm": 8.116076751113416,
+      "learning_rate": 3.031404958677686e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0452,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -9.5625,
+      "step": 1528
+    },
+    {
+      "epoch": 1.876073619631902,
+      "grad_norm": 7.632384073577,
+      "learning_rate": 3.028099173553719e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0225,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -10.1875,
+      "step": 1529
+    },
+    {
+      "epoch": 1.8773006134969326,
+      "grad_norm": 13.902306225751913,
+      "learning_rate": 3.024793388429752e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0604,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.408203125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.9375,
+      "step": 1530
+    },
+    {
+      "epoch": 1.8785276073619632,
+      "grad_norm": 15.95752061813539,
+      "learning_rate": 3.021487603305785e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0628,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": -0.201171875,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.5625,
+      "step": 1531
+    },
+    {
+      "epoch": 1.879754601226994,
+      "grad_norm": 13.945969752259758,
+      "learning_rate": 3.018181818181818e-07,
+      "logits/chosen": -0.3984375,
+      "logits/rejected": -0.57421875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -422.0,
+      "loss": 0.04,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -10.0625,
+      "step": 1532
+    },
+    {
+      "epoch": 1.8809815950920246,
+      "grad_norm": 9.516345960706976,
+      "learning_rate": 3.014876033057851e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0406,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.416015625,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -8.875,
+      "step": 1533
+    },
+    {
+      "epoch": 1.8822085889570552,
+      "grad_norm": 9.285807586877766,
+      "learning_rate": 3.0115702479338843e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0353,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1171875,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.8125,
+      "step": 1534
+    },
+    {
+      "epoch": 1.883435582822086,
+      "grad_norm": 16.084894756892087,
+      "learning_rate": 3.0082644628099175e-07,
+      "logits/chosen": -0.1513671875,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0633,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8828125,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.125,
+      "step": 1535
+    },
+    {
+      "epoch": 1.8846625766871166,
+      "grad_norm": 18.252488190346416,
+      "learning_rate": 3.00495867768595e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -450.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0635,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.25,
+      "step": 1536
+    },
+    {
+      "epoch": 1.8858895705521472,
+      "grad_norm": 6.3990194084358984,
+      "learning_rate": 3.0016528925619833e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0197,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5703125,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.75,
+      "step": 1537
+    },
+    {
+      "epoch": 1.887116564417178,
+      "grad_norm": 6.78081351935863,
+      "learning_rate": 2.9983471074380165e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0168,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.7890625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.375,
+      "step": 1538
+    },
+    {
+      "epoch": 1.8883435582822086,
+      "grad_norm": 8.405892255575104,
+      "learning_rate": 2.9950413223140496e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0352,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -8.8125,
+      "step": 1539
+    },
+    {
+      "epoch": 1.8895705521472392,
+      "grad_norm": 10.388061012954786,
+      "learning_rate": 2.9917355371900823e-07,
+      "logits/chosen": -0.39453125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -442.0,
+      "loss": 0.04,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.5,
+      "step": 1540
+    },
+    {
+      "epoch": 1.89079754601227,
+      "grad_norm": 11.04941546644624,
+      "learning_rate": 2.9884297520661155e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -380.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.50390625,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.5,
+      "step": 1541
+    },
+    {
+      "epoch": 1.8920245398773006,
+      "grad_norm": 9.143600142098991,
+      "learning_rate": 2.9851239669421486e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0244,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -9.8125,
+      "step": 1542
+    },
+    {
+      "epoch": 1.8932515337423312,
+      "grad_norm": 6.647581423330869,
+      "learning_rate": 2.981818181818182e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0238,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.61328125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.875,
+      "step": 1543
+    },
+    {
+      "epoch": 1.8944785276073621,
+      "grad_norm": 8.984495028391082,
+      "learning_rate": 2.978512396694215e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -426.0,
+      "loss": 0.034,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.78515625,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.75,
+      "step": 1544
+    },
+    {
+      "epoch": 1.8957055214723928,
+      "grad_norm": 8.685599093816752,
+      "learning_rate": 2.9752066115702476e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0309,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.040771484375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -8.9375,
+      "step": 1545
+    },
+    {
+      "epoch": 1.8969325153374235,
+      "grad_norm": 17.402219988439615,
+      "learning_rate": 2.971900826446281e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.46875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.375,
+      "step": 1546
+    },
+    {
+      "epoch": 1.8981595092024541,
+      "grad_norm": 14.691890087949803,
+      "learning_rate": 2.968595041322314e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -484.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0367,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.91796875,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -10.4375,
+      "step": 1547
+    },
+    {
+      "epoch": 1.8993865030674848,
+      "grad_norm": 28.990198997957158,
+      "learning_rate": 2.965289256198347e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0523,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.369140625,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.4375,
+      "step": 1548
+    },
+    {
+      "epoch": 1.9006134969325155,
+      "grad_norm": 9.555007462448875,
+      "learning_rate": 2.96198347107438e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0458,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.322265625,
+      "rewards/margins": 9.375,
+      "rewards/rejected": -9.0625,
+      "step": 1549
+    },
+    {
+      "epoch": 1.9018404907975461,
+      "grad_norm": 14.483826666374615,
+      "learning_rate": 2.958677685950413e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0474,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.423828125,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -10.125,
+      "step": 1550
+    },
+    {
+      "epoch": 1.9030674846625768,
+      "grad_norm": 13.761590699277777,
+      "learning_rate": 2.955371900826446e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -430.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0492,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.01031494140625,
+      "rewards/margins": 9.625,
+      "rewards/rejected": -9.625,
+      "step": 1551
+    },
+    {
+      "epoch": 1.9042944785276075,
+      "grad_norm": 13.384753485980882,
+      "learning_rate": 2.9520661157024793e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0606,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.15234375,
+      "rewards/margins": 8.6875,
+      "rewards/rejected": -8.5,
+      "step": 1552
+    },
+    {
+      "epoch": 1.9055214723926381,
+      "grad_norm": 14.733975887876666,
+      "learning_rate": 2.9487603305785125e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0354,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.3046875,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.5,
+      "step": 1553
+    },
+    {
+      "epoch": 1.9067484662576688,
+      "grad_norm": 13.95223783378209,
+      "learning_rate": 2.945454545454545e-07,
+      "logits/chosen": -0.369140625,
+      "logits/rejected": -0.6328125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0454,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.4140625,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -9.875,
+      "step": 1554
+    },
+    {
+      "epoch": 1.9079754601226995,
+      "grad_norm": 11.03554196573526,
+      "learning_rate": 2.9421487603305783e-07,
+      "logits/chosen": -0.390625,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0363,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.41015625,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.8125,
+      "step": 1555
+    },
+    {
+      "epoch": 1.9092024539877301,
+      "grad_norm": 9.08340310647596,
+      "learning_rate": 2.9388429752066115e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0364,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.64453125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.3125,
+      "step": 1556
+    },
+    {
+      "epoch": 1.9104294478527608,
+      "grad_norm": 12.6525315766818,
+      "learning_rate": 2.9355371900826447e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0477,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.330078125,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -9.0,
+      "step": 1557
+    },
+    {
+      "epoch": 1.9116564417177915,
+      "grad_norm": 9.233008765341841,
+      "learning_rate": 2.9322314049586773e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0246,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.53125,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -10.0625,
+      "step": 1558
+    },
+    {
+      "epoch": 1.9128834355828221,
+      "grad_norm": 16.102556892694622,
+      "learning_rate": 2.9289256198347105e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0634,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.30859375,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -9.0,
+      "step": 1559
+    },
+    {
+      "epoch": 1.9141104294478528,
+      "grad_norm": 8.41879847837055,
+      "learning_rate": 2.9256198347107437e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0306,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.59765625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -10.125,
+      "step": 1560
+    },
+    {
+      "epoch": 1.9153374233128835,
+      "grad_norm": 11.715145668756673,
+      "learning_rate": 2.922314049586777e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -452.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0351,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -10.0625,
+      "step": 1561
+    },
+    {
+      "epoch": 1.9165644171779141,
+      "grad_norm": 6.97755511576966,
+      "learning_rate": 2.9190082644628095e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -436.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0183,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -9.3125,
+      "step": 1562
+    },
+    {
+      "epoch": 1.9177914110429448,
+      "grad_norm": 7.472007517000815,
+      "learning_rate": 2.9157024793388427e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0251,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -9.625,
+      "step": 1563
+    },
+    {
+      "epoch": 1.9190184049079755,
+      "grad_norm": 8.89094542402821,
+      "learning_rate": 2.912396694214876e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0315,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.28125,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.1875,
+      "step": 1564
+    },
+    {
+      "epoch": 1.9202453987730062,
+      "grad_norm": 5.401693481934314,
+      "learning_rate": 2.909090909090909e-07,
+      "logits/chosen": -0.39453125,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -322.0,
+      "logps/rejected": -322.0,
+      "loss": 0.0252,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.5078125,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.5625,
+      "step": 1565
+    },
+    {
+      "epoch": 1.9214723926380368,
+      "grad_norm": 17.54721925631407,
+      "learning_rate": 2.905785123966942e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0475,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -9.5625,
+      "step": 1566
+    },
+    {
+      "epoch": 1.9226993865030675,
+      "grad_norm": 8.619832489981432,
+      "learning_rate": 2.902479338842975e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0383,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.408203125,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.75,
+      "step": 1567
+    },
+    {
+      "epoch": 1.9239263803680982,
+      "grad_norm": 9.440242433724526,
+      "learning_rate": 2.899173553719008e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -384.0,
+      "loss": 0.028,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -9.25,
+      "step": 1568
+    },
+    {
+      "epoch": 1.9251533742331288,
+      "grad_norm": 8.876395696240671,
+      "learning_rate": 2.895867768595041e-07,
+      "logits/chosen": -0.099609375,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0259,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.408203125,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.3125,
+      "step": 1569
+    },
+    {
+      "epoch": 1.9263803680981595,
+      "grad_norm": 7.999356547168104,
+      "learning_rate": 2.8925619834710744e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0328,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1796875,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.25,
+      "step": 1570
+    },
+    {
+      "epoch": 1.9276073619631902,
+      "grad_norm": 11.616155079550802,
+      "learning_rate": 2.889256198347107e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0441,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.1875,
+      "step": 1571
+    },
+    {
+      "epoch": 1.9288343558282208,
+      "grad_norm": 5.433830123431735,
+      "learning_rate": 2.88595041322314e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.5,
+      "logps/chosen": -412.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0176,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.419921875,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -10.1875,
+      "step": 1572
+    },
+    {
+      "epoch": 1.9300613496932515,
+      "grad_norm": 12.21034381437247,
+      "learning_rate": 2.8826446280991734e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0545,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.1826171875,
+      "rewards/margins": 9.25,
+      "rewards/rejected": -9.4375,
+      "step": 1573
+    },
+    {
+      "epoch": 1.9312883435582822,
+      "grad_norm": 13.499082687285052,
+      "learning_rate": 2.8793388429752066e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.40234375,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.125,
+      "step": 1574
+    },
+    {
+      "epoch": 1.9325153374233128,
+      "grad_norm": 15.391849130162866,
+      "learning_rate": 2.87603305785124e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -362.0,
+      "loss": 0.062,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.67578125,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.6875,
+      "step": 1575
+    },
+    {
+      "epoch": 1.9337423312883435,
+      "grad_norm": 8.29560522478457,
+      "learning_rate": 2.8727272727272724e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -364.0,
+      "loss": 0.026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.52734375,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -10.25,
+      "step": 1576
+    },
+    {
+      "epoch": 1.9349693251533742,
+      "grad_norm": 10.72329395494961,
+      "learning_rate": 2.8694214876033056e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0389,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.98046875,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.5,
+      "step": 1577
+    },
+    {
+      "epoch": 1.9361963190184048,
+      "grad_norm": 16.045872486554956,
+      "learning_rate": 2.866115702479339e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0492,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.0257568359375,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -10.0625,
+      "step": 1578
+    },
+    {
+      "epoch": 1.9374233128834355,
+      "grad_norm": 8.551245323989875,
+      "learning_rate": 2.862809917355372e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0344,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.279296875,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.75,
+      "step": 1579
+    },
+    {
+      "epoch": 1.9386503067484662,
+      "grad_norm": 6.503379535878091,
+      "learning_rate": 2.8595041322314046e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -444.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0201,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.828125,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.0625,
+      "step": 1580
+    },
+    {
+      "epoch": 1.9398773006134968,
+      "grad_norm": 8.183695473874987,
+      "learning_rate": 2.856198347107438e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0279,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.75,
+      "step": 1581
+    },
+    {
+      "epoch": 1.9411042944785275,
+      "grad_norm": 11.159282394202965,
+      "learning_rate": 2.852892561983471e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -330.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0554,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.1318359375,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -9.125,
+      "step": 1582
+    },
+    {
+      "epoch": 1.9423312883435582,
+      "grad_norm": 15.624281443403351,
+      "learning_rate": 2.849586776859504e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0559,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.92578125,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -9.6875,
+      "step": 1583
+    },
+    {
+      "epoch": 1.9435582822085888,
+      "grad_norm": 14.501392157008612,
+      "learning_rate": 2.8462809917355373e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -436.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0639,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.25,
+      "rewards/margins": 9.4375,
+      "rewards/rejected": -9.1875,
+      "step": 1584
+    },
+    {
+      "epoch": 1.9447852760736195,
+      "grad_norm": 12.361255357871427,
+      "learning_rate": 2.84297520661157e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0276,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.97265625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -10.625,
+      "step": 1585
+    },
+    {
+      "epoch": 1.9460122699386502,
+      "grad_norm": 12.44660865452298,
+      "learning_rate": 2.839669421487603e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0482,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.3125,
+      "step": 1586
+    },
+    {
+      "epoch": 1.9472392638036808,
+      "grad_norm": 10.155482386142491,
+      "learning_rate": 2.8363636363636363e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0395,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.06103515625,
+      "rewards/margins": 9.5,
+      "rewards/rejected": -9.4375,
+      "step": 1587
+    },
+    {
+      "epoch": 1.9484662576687115,
+      "grad_norm": 7.436422121571213,
+      "learning_rate": 2.8330578512396695e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0359,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.44921875,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.3125,
+      "step": 1588
+    },
+    {
+      "epoch": 1.9496932515337422,
+      "grad_norm": 8.76431349347538,
+      "learning_rate": 2.829752066115702e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0282,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2470703125,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -10.4375,
+      "step": 1589
+    },
+    {
+      "epoch": 1.9509202453987728,
+      "grad_norm": 7.189610220189287,
+      "learning_rate": 2.8264462809917353e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0227,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.83203125,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -9.9375,
+      "step": 1590
+    },
+    {
+      "epoch": 1.9521472392638037,
+      "grad_norm": 7.800363828306239,
+      "learning_rate": 2.8231404958677685e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0243,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1572265625,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -10.1875,
+      "step": 1591
+    },
+    {
+      "epoch": 1.9533742331288344,
+      "grad_norm": 6.9621963464144265,
+      "learning_rate": 2.8198347107438017e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0319,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.240234375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -9.9375,
+      "step": 1592
+    },
+    {
+      "epoch": 1.954601226993865,
+      "grad_norm": 9.330460459092231,
+      "learning_rate": 2.816528925619835e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0303,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.005218505859375,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -10.5625,
+      "step": 1593
+    },
+    {
+      "epoch": 1.9558282208588957,
+      "grad_norm": 13.631057553651138,
+      "learning_rate": 2.8132231404958675e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -328.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0628,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.26953125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.6875,
+      "step": 1594
+    },
+    {
+      "epoch": 1.9570552147239264,
+      "grad_norm": 14.520406844691742,
+      "learning_rate": 2.8099173553719007e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0438,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.287109375,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -10.8125,
+      "step": 1595
+    },
+    {
+      "epoch": 1.958282208588957,
+      "grad_norm": 7.279664367583889,
+      "learning_rate": 2.806611570247934e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0278,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.212890625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -10.25,
+      "step": 1596
+    },
+    {
+      "epoch": 1.9595092024539877,
+      "grad_norm": 11.848236709613104,
+      "learning_rate": 2.803305785123967e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0332,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.349609375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.5,
+      "step": 1597
+    },
+    {
+      "epoch": 1.9607361963190184,
+      "grad_norm": 20.200578986204675,
+      "learning_rate": 2.7999999999999997e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0706,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.28515625,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -10.1875,
+      "step": 1598
+    },
+    {
+      "epoch": 1.961963190184049,
+      "grad_norm": 6.9742046512408855,
+      "learning_rate": 2.796694214876033e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0277,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1044921875,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -10.125,
+      "step": 1599
+    },
+    {
+      "epoch": 1.9631901840490797,
+      "grad_norm": 12.052751187248274,
+      "learning_rate": 2.793388429752066e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0386,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.2041015625,
+      "rewards/margins": 10.0625,
+      "rewards/rejected": -9.875,
+      "step": 1600
+    },
+    {
+      "epoch": 1.9644171779141104,
+      "grad_norm": 10.394378430820012,
+      "learning_rate": 2.790082644628099e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.25,
+      "logps/chosen": -342.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0494,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.41796875,
+      "rewards/margins": 9.0625,
+      "rewards/rejected": -9.5,
+      "step": 1601
+    },
+    {
+      "epoch": 1.965644171779141,
+      "grad_norm": 11.67542858875569,
+      "learning_rate": 2.7867768595041324e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -438.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0411,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -9.5,
+      "step": 1602
+    },
+    {
+      "epoch": 1.9668711656441717,
+      "grad_norm": 9.973844708079088,
+      "learning_rate": 2.783471074380165e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -468.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0408,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.08544921875,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -10.6875,
+      "step": 1603
+    },
+    {
+      "epoch": 1.9680981595092024,
+      "grad_norm": 12.506574586450439,
+      "learning_rate": 2.780165289256198e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0453,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.46484375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -9.0625,
+      "step": 1604
+    },
+    {
+      "epoch": 1.969325153374233,
+      "grad_norm": 14.4298856829564,
+      "learning_rate": 2.7768595041322314e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -428.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0627,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": -0.171875,
+      "rewards/margins": 9.75,
+      "rewards/rejected": -9.9375,
+      "step": 1605
+    },
+    {
+      "epoch": 1.9705521472392638,
+      "grad_norm": 10.928070857156225,
+      "learning_rate": 2.7735537190082645e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -414.0,
+      "loss": 0.032,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.8984375,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -10.1875,
+      "step": 1606
+    },
+    {
+      "epoch": 1.9717791411042946,
+      "grad_norm": 10.048807722409952,
+      "learning_rate": 2.770247933884297e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.470703125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -344.0,
+      "loss": 0.036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.546875,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.875,
+      "step": 1607
+    },
+    {
+      "epoch": 1.9730061349693253,
+      "grad_norm": 9.370791095894255,
+      "learning_rate": 2.7669421487603304e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0296,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.3671875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -10.0,
+      "step": 1608
+    },
+    {
+      "epoch": 1.974233128834356,
+      "grad_norm": 13.452578592993975,
+      "learning_rate": 2.7636363636363635e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0525,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.263671875,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -10.4375,
+      "step": 1609
+    },
+    {
+      "epoch": 1.9754601226993866,
+      "grad_norm": 17.64298790831311,
+      "learning_rate": 2.7603305785123967e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0446,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.08056640625,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.875,
+      "step": 1610
+    },
+    {
+      "epoch": 1.9766871165644173,
+      "grad_norm": 11.712184783781975,
+      "learning_rate": 2.7570247933884294e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -326.0,
+      "logps/rejected": -322.0,
+      "loss": 0.0566,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.55859375,
+      "rewards/margins": 8.75,
+      "rewards/rejected": -9.3125,
+      "step": 1611
+    },
+    {
+      "epoch": 1.977914110429448,
+      "grad_norm": 11.429670120969462,
+      "learning_rate": 2.7537190082644626e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -440.0,
+      "loss": 0.0371,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.6875,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -10.125,
+      "step": 1612
+    },
+    {
+      "epoch": 1.9791411042944786,
+      "grad_norm": 9.475634561082307,
+      "learning_rate": 2.7504132231404957e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -440.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0259,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.380859375,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -10.9375,
+      "step": 1613
+    },
+    {
+      "epoch": 1.9803680981595093,
+      "grad_norm": 15.195294834798235,
+      "learning_rate": 2.747107438016529e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0612,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.287109375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.5625,
+      "step": 1614
+    },
+    {
+      "epoch": 1.98159509202454,
+      "grad_norm": 17.650174474381533,
+      "learning_rate": 2.743801652892562e-07,
+      "logits/chosen": -0.375,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0668,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 0.25390625,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -10.6875,
+      "step": 1615
+    },
+    {
+      "epoch": 1.9828220858895707,
+      "grad_norm": 15.324230092096657,
+      "learning_rate": 2.7404958677685947e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0512,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.4609375,
+      "rewards/margins": 9.5625,
+      "rewards/rejected": -10.0,
+      "step": 1616
+    },
+    {
+      "epoch": 1.9840490797546013,
+      "grad_norm": 19.69575258656083,
+      "learning_rate": 2.737190082644628e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -356.0,
+      "loss": 0.06,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.36328125,
+      "rewards/margins": 9.0,
+      "rewards/rejected": -9.3125,
+      "step": 1617
+    },
+    {
+      "epoch": 1.985276073619632,
+      "grad_norm": 8.216862123400189,
+      "learning_rate": 2.733884297520661e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0363,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.0419921875,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.875,
+      "step": 1618
+    },
+    {
+      "epoch": 1.9865030674846627,
+      "grad_norm": 10.46369822241489,
+      "learning_rate": 2.730578512396694e-07,
+      "logits/chosen": -0.330078125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0435,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.20703125,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -10.25,
+      "step": 1619
+    },
+    {
+      "epoch": 1.9877300613496933,
+      "grad_norm": 12.60171916853773,
+      "learning_rate": 2.727272727272727e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -362.0,
+      "loss": 0.033,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.00213623046875,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -10.3125,
+      "step": 1620
+    },
+    {
+      "epoch": 1.988957055214724,
+      "grad_norm": 8.365282012776243,
+      "learning_rate": 2.72396694214876e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0317,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.3359375,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -10.25,
+      "step": 1621
+    },
+    {
+      "epoch": 1.9901840490797547,
+      "grad_norm": 11.044384202131916,
+      "learning_rate": 2.720661157024793e-07,
+      "logits/chosen": -0.455078125,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0438,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.1552734375,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -10.5,
+      "step": 1622
+    },
+    {
+      "epoch": 1.9914110429447853,
+      "grad_norm": 8.905193463781423,
+      "learning_rate": 2.7173553719008264e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0235,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.36328125,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -10.375,
+      "step": 1623
+    },
+    {
+      "epoch": 1.992638036809816,
+      "grad_norm": 6.811861007896774,
+      "learning_rate": 2.7140495867768596e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0255,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -10.375,
+      "step": 1624
+    },
+    {
+      "epoch": 1.9938650306748467,
+      "grad_norm": 10.024309031384108,
+      "learning_rate": 2.710743801652892e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -442.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0316,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -10.0,
+      "step": 1625
+    },
+    {
+      "epoch": 1.9950920245398773,
+      "grad_norm": 18.695693680609278,
+      "learning_rate": 2.7074380165289254e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0647,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1953125,
+      "rewards/margins": 9.9375,
+      "rewards/rejected": -9.75,
+      "step": 1626
+    },
+    {
+      "epoch": 1.996319018404908,
+      "grad_norm": 10.936992214829806,
+      "learning_rate": 2.7041322314049586e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0422,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.58984375,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -10.0,
+      "step": 1627
+    },
+    {
+      "epoch": 1.9975460122699387,
+      "grad_norm": 9.584409085330941,
+      "learning_rate": 2.700826446280992e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0422,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.1875,
+      "step": 1628
+    },
+    {
+      "epoch": 1.9987730061349693,
+      "grad_norm": 13.453298121808412,
+      "learning_rate": 2.6975206611570244e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0514,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": 0.2060546875,
+      "rewards/margins": 10.0,
+      "rewards/rejected": -9.8125,
+      "step": 1629
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 11.391241637712861,
+      "learning_rate": 2.6942148760330576e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0366,
+      "rewards/accuracies": 0.9915966987609863,
+      "rewards/chosen": 0.4140625,
+      "rewards/margins": 11.125,
+      "rewards/rejected": -10.6875,
+      "step": 1630
+    },
+    {
+      "epoch": 2.0012269938650307,
+      "grad_norm": 1.0056092333572968,
+      "learning_rate": 2.690909090909091e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -430.0,
+      "logps/rejected": -378.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -10.375,
+      "step": 1631
+    },
+    {
+      "epoch": 2.0024539877300613,
+      "grad_norm": 1.6634039758609083,
+      "learning_rate": 2.687603305785124e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -10.125,
+      "step": 1632
+    },
+    {
+      "epoch": 2.003680981595092,
+      "grad_norm": 4.503480532585863,
+      "learning_rate": 2.684297520661157e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -464.0,
+      "logps/rejected": -384.0,
+      "loss": 0.019,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -9.8125,
+      "step": 1633
+    },
+    {
+      "epoch": 2.0049079754601227,
+      "grad_norm": 2.6023603762729968,
+      "learning_rate": 2.68099173553719e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0092,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.41015625,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -10.0,
+      "step": 1634
+    },
+    {
+      "epoch": 2.0061349693251533,
+      "grad_norm": 1.8319060305008925,
+      "learning_rate": 2.677685950413223e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -10.5625,
+      "step": 1635
+    },
+    {
+      "epoch": 2.007361963190184,
+      "grad_norm": 3.17239320052001,
+      "learning_rate": 2.674380165289256e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -376.0,
+      "loss": 0.007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.63671875,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -10.5,
+      "step": 1636
+    },
+    {
+      "epoch": 2.0085889570552147,
+      "grad_norm": 1.5637038220704362,
+      "learning_rate": 2.6710743801652893e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -334.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0056,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.11083984375,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -10.0625,
+      "step": 1637
+    },
+    {
+      "epoch": 2.0098159509202453,
+      "grad_norm": 1.5615192510805758,
+      "learning_rate": 2.667768595041322e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0063,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.97265625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -10.5625,
+      "step": 1638
+    },
+    {
+      "epoch": 2.011042944785276,
+      "grad_norm": 1.6057410482160726,
+      "learning_rate": 2.664462809917355e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -346.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -9.875,
+      "step": 1639
+    },
+    {
+      "epoch": 2.0122699386503067,
+      "grad_norm": 1.6911694052858361,
+      "learning_rate": 2.6611570247933883e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -9.625,
+      "step": 1640
+    },
+    {
+      "epoch": 2.0134969325153373,
+      "grad_norm": 2.0273278923099127,
+      "learning_rate": 2.6578512396694215e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.8828125,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -9.5,
+      "step": 1641
+    },
+    {
+      "epoch": 2.014723926380368,
+      "grad_norm": 2.268462880120689,
+      "learning_rate": 2.6545454545454547e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0121,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -10.0,
+      "step": 1642
+    },
+    {
+      "epoch": 2.0159509202453987,
+      "grad_norm": 1.7124682486644203,
+      "learning_rate": 2.6512396694214873e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -334.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0065,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1376953125,
+      "rewards/margins": 9.8125,
+      "rewards/rejected": -9.6875,
+      "step": 1643
+    },
+    {
+      "epoch": 2.0171779141104293,
+      "grad_norm": 3.4517000116531475,
+      "learning_rate": 2.6479338842975205e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0108,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -9.5625,
+      "step": 1644
+    },
+    {
+      "epoch": 2.01840490797546,
+      "grad_norm": 1.6132270494784144,
+      "learning_rate": 2.6446280991735537e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0073,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -9.9375,
+      "step": 1645
+    },
+    {
+      "epoch": 2.0196319018404907,
+      "grad_norm": 1.9044833617268588,
+      "learning_rate": 2.641322314049587e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0062,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5546875,
+      "rewards/margins": 9.875,
+      "rewards/rejected": -9.3125,
+      "step": 1646
+    },
+    {
+      "epoch": 2.0208588957055214,
+      "grad_norm": 1.8869849147750222,
+      "learning_rate": 2.6380165289256195e-07,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.486328125,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -10.5625,
+      "step": 1647
+    },
+    {
+      "epoch": 2.022085889570552,
+      "grad_norm": 1.9916590971460693,
+      "learning_rate": 2.6347107438016527e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0125,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.490234375,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.0625,
+      "step": 1648
+    },
+    {
+      "epoch": 2.0233128834355827,
+      "grad_norm": 1.4707935848394005,
+      "learning_rate": 2.631404958677686e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -9.8125,
+      "step": 1649
+    },
+    {
+      "epoch": 2.0245398773006134,
+      "grad_norm": 2.3373611075810574,
+      "learning_rate": 2.628099173553719e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -340.0,
+      "loss": 0.0122,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.515625,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -9.6875,
+      "step": 1650
+    },
+    {
+      "epoch": 2.025766871165644,
+      "grad_norm": 1.978858737441041,
+      "learning_rate": 2.6247933884297517e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -410.0,
+      "loss": 0.005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -10.1875,
+      "step": 1651
+    },
+    {
+      "epoch": 2.0269938650306747,
+      "grad_norm": 1.7855183239701438,
+      "learning_rate": 2.621487603305785e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -340.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0066,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.65234375,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.875,
+      "step": 1652
+    },
+    {
+      "epoch": 2.0282208588957054,
+      "grad_norm": 1.5585444997367146,
+      "learning_rate": 2.618181818181818e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0118,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.82421875,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -10.1875,
+      "step": 1653
+    },
+    {
+      "epoch": 2.029447852760736,
+      "grad_norm": 4.17907207360699,
+      "learning_rate": 2.614876033057851e-07,
+      "logits/chosen": -0.060546875,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0132,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.72265625,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -9.8125,
+      "step": 1654
+    },
+    {
+      "epoch": 2.0306748466257667,
+      "grad_norm": 2.281718883715343,
+      "learning_rate": 2.6115702479338844e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0075,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -10.6875,
+      "step": 1655
+    },
+    {
+      "epoch": 2.0319018404907974,
+      "grad_norm": 1.333127158804313,
+      "learning_rate": 2.608264462809917e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -444.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.63671875,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -10.375,
+      "step": 1656
+    },
+    {
+      "epoch": 2.033128834355828,
+      "grad_norm": 2.2517421914913127,
+      "learning_rate": 2.60495867768595e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0073,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.78125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -9.625,
+      "step": 1657
+    },
+    {
+      "epoch": 2.0343558282208587,
+      "grad_norm": 6.274063932592078,
+      "learning_rate": 2.6016528925619834e-07,
+      "logits/chosen": -0.24609375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0194,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.357421875,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.5,
+      "step": 1658
+    },
+    {
+      "epoch": 2.0355828220858894,
+      "grad_norm": 1.5458804787455829,
+      "learning_rate": 2.5983471074380166e-07,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0888671875,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.5,
+      "step": 1659
+    },
+    {
+      "epoch": 2.03680981595092,
+      "grad_norm": 1.948609467781341,
+      "learning_rate": 2.595041322314049e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0069,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.255859375,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -10.1875,
+      "step": 1660
+    },
+    {
+      "epoch": 2.038036809815951,
+      "grad_norm": 1.3640208870503872,
+      "learning_rate": 2.5917355371900824e-07,
+      "logits/chosen": -0.384765625,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2041015625,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -11.125,
+      "step": 1661
+    },
+    {
+      "epoch": 2.039263803680982,
+      "grad_norm": 1.7729903326221843,
+      "learning_rate": 2.5884297520661156e-07,
+      "logits/chosen": -0.37890625,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -466.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -11.1875,
+      "step": 1662
+    },
+    {
+      "epoch": 2.0404907975460125,
+      "grad_norm": 2.0189355451807476,
+      "learning_rate": 2.585123966942149e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.51953125,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -10.9375,
+      "step": 1663
+    },
+    {
+      "epoch": 2.041717791411043,
+      "grad_norm": 1.1835273153892374,
+      "learning_rate": 2.581818181818182e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.427734375,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -11.0,
+      "step": 1664
+    },
+    {
+      "epoch": 2.042944785276074,
+      "grad_norm": 2.7384556051554267,
+      "learning_rate": 2.5785123966942146e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0095,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.0014801025390625,
+      "rewards/margins": 10.1875,
+      "rewards/rejected": -10.1875,
+      "step": 1665
+    },
+    {
+      "epoch": 2.0441717791411045,
+      "grad_norm": 4.19437613372316,
+      "learning_rate": 2.575206611570248e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0117,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -10.875,
+      "step": 1666
+    },
+    {
+      "epoch": 2.045398773006135,
+      "grad_norm": 2.410224884231637,
+      "learning_rate": 2.571900826446281e-07,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0116,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.392578125,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -11.125,
+      "step": 1667
+    },
+    {
+      "epoch": 2.046625766871166,
+      "grad_norm": 0.8317861949740706,
+      "learning_rate": 2.568595041322314e-07,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -366.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0024,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -10.9375,
+      "step": 1668
+    },
+    {
+      "epoch": 2.0478527607361965,
+      "grad_norm": 1.5732335806397932,
+      "learning_rate": 2.565289256198347e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0109,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.29296875,
+      "rewards/margins": 10.375,
+      "rewards/rejected": -10.125,
+      "step": 1669
+    },
+    {
+      "epoch": 2.049079754601227,
+      "grad_norm": 2.296600003852521,
+      "learning_rate": 2.56198347107438e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0061,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -10.875,
+      "step": 1670
+    },
+    {
+      "epoch": 2.050306748466258,
+      "grad_norm": 1.1800097595662804,
+      "learning_rate": 2.558677685950413e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.9921875,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -10.8125,
+      "step": 1671
+    },
+    {
+      "epoch": 2.0515337423312885,
+      "grad_norm": 1.4722735997256955,
+      "learning_rate": 2.5553719008264463e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.90625,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -10.8125,
+      "step": 1672
+    },
+    {
+      "epoch": 2.052760736196319,
+      "grad_norm": 2.0056968603212373,
+      "learning_rate": 2.5520661157024795e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -374.0,
+      "loss": 0.007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.51171875,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -10.125,
+      "step": 1673
+    },
+    {
+      "epoch": 2.05398773006135,
+      "grad_norm": 1.5023788016839823,
+      "learning_rate": 2.548760330578512e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.6015625,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -11.375,
+      "step": 1674
+    },
+    {
+      "epoch": 2.0552147239263805,
+      "grad_norm": 2.343189055479409,
+      "learning_rate": 2.5454545454545453e-07,
+      "logits/chosen": -0.31640625,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -440.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0068,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.75,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -10.875,
+      "step": 1675
+    },
+    {
+      "epoch": 2.056441717791411,
+      "grad_norm": 2.029437741658446,
+      "learning_rate": 2.5421487603305785e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.4375,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -10.875,
+      "step": 1676
+    },
+    {
+      "epoch": 2.057668711656442,
+      "grad_norm": 1.67014589884543,
+      "learning_rate": 2.5388429752066116e-07,
+      "logits/chosen": -0.4140625,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0092,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.75,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -12.125,
+      "step": 1677
+    },
+    {
+      "epoch": 2.0588957055214725,
+      "grad_norm": 1.661274681544124,
+      "learning_rate": 2.5355371900826443e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.392578125,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -11.3125,
+      "step": 1678
+    },
+    {
+      "epoch": 2.060122699386503,
+      "grad_norm": 1.3319482224823802,
+      "learning_rate": 2.5322314049586775e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -302.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.369140625,
+      "rewards/margins": 10.25,
+      "rewards/rejected": -10.625,
+      "step": 1679
+    },
+    {
+      "epoch": 2.061349693251534,
+      "grad_norm": 1.7002471004141169,
+      "learning_rate": 2.5289256198347106e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0057,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.349609375,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -10.5625,
+      "step": 1680
+    },
+    {
+      "epoch": 2.0625766871165645,
+      "grad_norm": 1.3165764400408078,
+      "learning_rate": 2.525619834710744e-07,
+      "logits/chosen": -0.34765625,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -414.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.953125,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -11.125,
+      "step": 1681
+    },
+    {
+      "epoch": 2.063803680981595,
+      "grad_norm": 1.447457988554957,
+      "learning_rate": 2.522314049586777e-07,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.181640625,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -10.75,
+      "step": 1682
+    },
+    {
+      "epoch": 2.065030674846626,
+      "grad_norm": 1.9179967141243504,
+      "learning_rate": 2.5190082644628096e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.02734375,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -10.625,
+      "step": 1683
+    },
+    {
+      "epoch": 2.0662576687116565,
+      "grad_norm": 3.4613367299669435,
+      "learning_rate": 2.515702479338843e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0095,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.71484375,
+      "rewards/margins": 9.6875,
+      "rewards/rejected": -10.4375,
+      "step": 1684
+    },
+    {
+      "epoch": 2.067484662576687,
+      "grad_norm": 2.899751055899642,
+      "learning_rate": 2.512396694214876e-07,
+      "logits/chosen": -0.1357421875,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0108,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.353515625,
+      "rewards/margins": 10.125,
+      "rewards/rejected": -10.5,
+      "step": 1685
+    },
+    {
+      "epoch": 2.068711656441718,
+      "grad_norm": 1.3385737500657435,
+      "learning_rate": 2.509090909090909e-07,
+      "logits/chosen": -0.392578125,
+      "logits/rejected": -0.58203125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.017578125,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -10.9375,
+      "step": 1686
+    },
+    {
+      "epoch": 2.0699386503067485,
+      "grad_norm": 2.4750726240620966,
+      "learning_rate": 2.505785123966942e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.49609375,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -11.125,
+      "step": 1687
+    },
+    {
+      "epoch": 2.071165644171779,
+      "grad_norm": 1.863240161619954,
+      "learning_rate": 2.502479338842975e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2265625,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -11.25,
+      "step": 1688
+    },
+    {
+      "epoch": 2.07239263803681,
+      "grad_norm": 2.4607621842494383,
+      "learning_rate": 2.499173553719008e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0072,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.4921875,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -10.875,
+      "step": 1689
+    },
+    {
+      "epoch": 2.0736196319018405,
+      "grad_norm": 1.4786219841231059,
+      "learning_rate": 2.4958677685950414e-07,
+      "logits/chosen": -0.365234375,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.47265625,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -11.5,
+      "step": 1690
+    },
+    {
+      "epoch": 2.074846625766871,
+      "grad_norm": 1.7097590583119724,
+      "learning_rate": 2.492561983471074e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.19140625,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -10.875,
+      "step": 1691
+    },
+    {
+      "epoch": 2.076073619631902,
+      "grad_norm": 2.452796956426982,
+      "learning_rate": 2.489256198347107e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0092,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.388671875,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -10.875,
+      "step": 1692
+    },
+    {
+      "epoch": 2.0773006134969325,
+      "grad_norm": 1.2276222258825435,
+      "learning_rate": 2.4859504132231404e-07,
+      "logits/chosen": -0.3515625,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -430.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.0634765625,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -11.6875,
+      "step": 1693
+    },
+    {
+      "epoch": 2.078527607361963,
+      "grad_norm": 1.6031739418209205,
+      "learning_rate": 2.4826446280991735e-07,
+      "logits/chosen": -0.40625,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0099,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.25390625,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -10.8125,
+      "step": 1694
+    },
+    {
+      "epoch": 2.079754601226994,
+      "grad_norm": 2.4277801118685627,
+      "learning_rate": 2.4793388429752067e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0121,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.498046875,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -11.0,
+      "step": 1695
+    },
+    {
+      "epoch": 2.0809815950920245,
+      "grad_norm": 3.397787671749636,
+      "learning_rate": 2.4760330578512394e-07,
+      "logits/chosen": -0.41015625,
+      "logits/rejected": -0.63671875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.203125,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -11.875,
+      "step": 1696
+    },
+    {
+      "epoch": 2.082208588957055,
+      "grad_norm": 1.7000818711464039,
+      "learning_rate": 2.4727272727272725e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.30859375,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -11.5625,
+      "step": 1697
+    },
+    {
+      "epoch": 2.083435582822086,
+      "grad_norm": 3.9971262968351615,
+      "learning_rate": 2.4694214876033057e-07,
+      "logits/chosen": -0.341796875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0069,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5703125,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.0,
+      "step": 1698
+    },
+    {
+      "epoch": 2.0846625766871165,
+      "grad_norm": 2.0755351799905406,
+      "learning_rate": 2.466115702479339e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -462.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.6875,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -11.0625,
+      "step": 1699
+    },
+    {
+      "epoch": 2.085889570552147,
+      "grad_norm": 1.8613638158791848,
+      "learning_rate": 2.4628099173553715e-07,
+      "logits/chosen": -0.2001953125,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0063,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1611328125,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -10.4375,
+      "step": 1700
+    },
+    {
+      "epoch": 2.087116564417178,
+      "grad_norm": 2.017916099112244,
+      "learning_rate": 2.4595041322314047e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.66796875,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -11.625,
+      "step": 1701
+    },
+    {
+      "epoch": 2.0883435582822085,
+      "grad_norm": 1.5190053044495557,
+      "learning_rate": 2.456198347107438e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -434.0,
+      "logps/rejected": -404.0,
+      "loss": 0.009,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.421875,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -12.25,
+      "step": 1702
+    },
+    {
+      "epoch": 2.089570552147239,
+      "grad_norm": 1.3611732393061429,
+      "learning_rate": 2.452892561983471e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0299072265625,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -11.875,
+      "step": 1703
+    },
+    {
+      "epoch": 2.09079754601227,
+      "grad_norm": 2.201142044064524,
+      "learning_rate": 2.449586776859504e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -460.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 13.9375,
+      "rewards/rejected": -12.75,
+      "step": 1704
+    },
+    {
+      "epoch": 2.0920245398773005,
+      "grad_norm": 2.3208253690657057,
+      "learning_rate": 2.446280991735537e-07,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.490234375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.091796875,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -11.6875,
+      "step": 1705
+    },
+    {
+      "epoch": 2.093251533742331,
+      "grad_norm": 2.8234711397459953,
+      "learning_rate": 2.44297520661157e-07,
+      "logits/chosen": -0.37109375,
+      "logits/rejected": -0.60546875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0061,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.58984375,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -11.375,
+      "step": 1706
+    },
+    {
+      "epoch": 2.094478527607362,
+      "grad_norm": 1.5991143629396039,
+      "learning_rate": 2.439669421487603e-07,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2451171875,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -11.625,
+      "step": 1707
+    },
+    {
+      "epoch": 2.0957055214723925,
+      "grad_norm": 5.857155543388813,
+      "learning_rate": 2.4363636363636364e-07,
+      "logits/chosen": -0.39453125,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0071,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.048828125,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.375,
+      "step": 1708
+    },
+    {
+      "epoch": 2.096932515337423,
+      "grad_norm": 2.3029106551896272,
+      "learning_rate": 2.433057851239669e-07,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0074,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.0130615234375,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -11.6875,
+      "step": 1709
+    },
+    {
+      "epoch": 2.098159509202454,
+      "grad_norm": 2.102525855847367,
+      "learning_rate": 2.429752066115702e-07,
+      "logits/chosen": -0.41796875,
+      "logits/rejected": -0.62109375,
+      "logps/chosen": -472.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0089,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.330078125,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -12.0,
+      "step": 1710
+    },
+    {
+      "epoch": 2.0993865030674845,
+      "grad_norm": 3.345482925647388,
+      "learning_rate": 2.4264462809917354e-07,
+      "logits/chosen": -0.353515625,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2275390625,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -11.4375,
+      "step": 1711
+    },
+    {
+      "epoch": 2.100613496932515,
+      "grad_norm": 2.1821746267796627,
+      "learning_rate": 2.4231404958677686e-07,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.330078125,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -11.5625,
+      "step": 1712
+    },
+    {
+      "epoch": 2.101840490797546,
+      "grad_norm": 0.9947645133297353,
+      "learning_rate": 2.419834710743802e-07,
+      "logits/chosen": -0.357421875,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0136,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 0.365234375,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -11.75,
+      "step": 1713
+    },
+    {
+      "epoch": 2.1030674846625765,
+      "grad_norm": 1.4638206192490053,
+      "learning_rate": 2.4165289256198344e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -350.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.06689453125,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -11.0,
+      "step": 1714
+    },
+    {
+      "epoch": 2.104294478527607,
+      "grad_norm": 1.9832721541155545,
+      "learning_rate": 2.4132231404958676e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -396.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.01904296875,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -11.75,
+      "step": 1715
+    },
+    {
+      "epoch": 2.105521472392638,
+      "grad_norm": 3.8450353097214407,
+      "learning_rate": 2.409917355371901e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -394.0,
+      "loss": 0.009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.3046875,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -10.625,
+      "step": 1716
+    },
+    {
+      "epoch": 2.1067484662576685,
+      "grad_norm": 1.3438587089970906,
+      "learning_rate": 2.406611570247934e-07,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.54296875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0038,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.22265625,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -11.8125,
+      "step": 1717
+    },
+    {
+      "epoch": 2.107975460122699,
+      "grad_norm": 1.4243801375622738,
+      "learning_rate": 2.4033057851239666e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0091,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.1298828125,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -11.0625,
+      "step": 1718
+    },
+    {
+      "epoch": 2.10920245398773,
+      "grad_norm": 1.464829451908975,
+      "learning_rate": 2.4e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -440.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0136,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -11.75,
+      "step": 1719
+    },
+    {
+      "epoch": 2.1104294478527605,
+      "grad_norm": 1.5396085980892982,
+      "learning_rate": 2.396694214876033e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.06640625,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -11.125,
+      "step": 1720
+    },
+    {
+      "epoch": 2.111656441717791,
+      "grad_norm": 2.3557517273261213,
+      "learning_rate": 2.393388429752066e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.55078125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0058,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -12.375,
+      "step": 1721
+    },
+    {
+      "epoch": 2.112883435582822,
+      "grad_norm": 3.2976605852634036,
+      "learning_rate": 2.3900826446280993e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -380.0,
+      "loss": 0.01,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.71484375,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -11.75,
+      "step": 1722
+    },
+    {
+      "epoch": 2.1141104294478525,
+      "grad_norm": 1.202939252606192,
+      "learning_rate": 2.386776859504132e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.54296875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.08154296875,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -11.125,
+      "step": 1723
+    },
+    {
+      "epoch": 2.1153374233128837,
+      "grad_norm": 2.356805196113207,
+      "learning_rate": 2.3834710743801651e-07,
+      "logits/chosen": -0.361328125,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -346.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.259765625,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -11.25,
+      "step": 1724
+    },
+    {
+      "epoch": 2.116564417177914,
+      "grad_norm": 1.208111474668117,
+      "learning_rate": 2.3801652892561983e-07,
+      "logits/chosen": -0.345703125,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1015625,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -11.9375,
+      "step": 1725
+    },
+    {
+      "epoch": 2.117791411042945,
+      "grad_norm": 1.7415518072746483,
+      "learning_rate": 2.3768595041322315e-07,
+      "logits/chosen": -0.4375,
+      "logits/rejected": -0.6171875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0109,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.58984375,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.1875,
+      "step": 1726
+    },
+    {
+      "epoch": 2.1190184049079757,
+      "grad_norm": 1.0683994638090692,
+      "learning_rate": 2.3735537190082641e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0025,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 13.4375,
+      "rewards/rejected": -12.4375,
+      "step": 1727
+    },
+    {
+      "epoch": 2.1202453987730063,
+      "grad_norm": 2.158821333101882,
+      "learning_rate": 2.3702479338842973e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -434.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0079,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.224609375,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -11.75,
+      "step": 1728
+    },
+    {
+      "epoch": 2.121472392638037,
+      "grad_norm": 1.49660767422767,
+      "learning_rate": 2.3669421487603305e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.3203125,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -11.3125,
+      "step": 1729
+    },
+    {
+      "epoch": 2.1226993865030677,
+      "grad_norm": 1.8550719153721926,
+      "learning_rate": 2.3636363636363637e-07,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -404.0,
+      "loss": 0.007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.015625,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -12.0625,
+      "step": 1730
+    },
+    {
+      "epoch": 2.1239263803680983,
+      "grad_norm": 1.692477865271817,
+      "learning_rate": 2.3603305785123969e-07,
+      "logits/chosen": -0.34765625,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.30078125,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.875,
+      "step": 1731
+    },
+    {
+      "epoch": 2.125153374233129,
+      "grad_norm": 1.1159030063879751,
+      "learning_rate": 2.3570247933884295e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.515625,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.3125,
+      "step": 1732
+    },
+    {
+      "epoch": 2.1263803680981597,
+      "grad_norm": 1.3850434795304294,
+      "learning_rate": 2.3537190082644627e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.6484375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -11.5625,
+      "step": 1733
+    },
+    {
+      "epoch": 2.1276073619631903,
+      "grad_norm": 1.698094413995771,
+      "learning_rate": 2.3504132231404959e-07,
+      "logits/chosen": -0.341796875,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -432.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.224609375,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -12.875,
+      "step": 1734
+    },
+    {
+      "epoch": 2.128834355828221,
+      "grad_norm": 1.5979568554781107,
+      "learning_rate": 2.347107438016529e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.59765625,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -12.125,
+      "step": 1735
+    },
+    {
+      "epoch": 2.1300613496932517,
+      "grad_norm": 2.7088652602034635,
+      "learning_rate": 2.3438016528925617e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -418.0,
+      "loss": 0.007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6015625,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -11.4375,
+      "step": 1736
+    },
+    {
+      "epoch": 2.1312883435582823,
+      "grad_norm": 2.1884142647436766,
+      "learning_rate": 2.3404958677685949e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.057861328125,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -11.875,
+      "step": 1737
+    },
+    {
+      "epoch": 2.132515337423313,
+      "grad_norm": 1.9976390353669047,
+      "learning_rate": 2.337190082644628e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0102,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.462890625,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.4375,
+      "step": 1738
+    },
+    {
+      "epoch": 2.1337423312883437,
+      "grad_norm": 2.635940204938581,
+      "learning_rate": 2.3338842975206612e-07,
+      "logits/chosen": -0.322265625,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0082,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.345703125,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.3125,
+      "step": 1739
+    },
+    {
+      "epoch": 2.1349693251533743,
+      "grad_norm": 1.3870343898827924,
+      "learning_rate": 2.3305785123966939e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0092,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.15625,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -11.875,
+      "step": 1740
+    },
+    {
+      "epoch": 2.136196319018405,
+      "grad_norm": 5.607847284446492,
+      "learning_rate": 2.327272727272727e-07,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0081,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.30859375,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.0,
+      "step": 1741
+    },
+    {
+      "epoch": 2.1374233128834357,
+      "grad_norm": 1.9535733016897885,
+      "learning_rate": 2.3239669421487602e-07,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.470703125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1025390625,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -11.8125,
+      "step": 1742
+    },
+    {
+      "epoch": 2.1386503067484663,
+      "grad_norm": 1.4899587259798763,
+      "learning_rate": 2.3206611570247934e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0096,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.33203125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -11.9375,
+      "step": 1743
+    },
+    {
+      "epoch": 2.139877300613497,
+      "grad_norm": 2.624189401140415,
+      "learning_rate": 2.3173553719008266e-07,
+      "logits/chosen": -0.41796875,
+      "logits/rejected": -0.66796875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.431640625,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.0625,
+      "step": 1744
+    },
+    {
+      "epoch": 2.1411042944785277,
+      "grad_norm": 1.480829103043516,
+      "learning_rate": 2.3140495867768592e-07,
+      "logits/chosen": -0.3671875,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.419921875,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.625,
+      "step": 1745
+    },
+    {
+      "epoch": 2.1423312883435583,
+      "grad_norm": 1.8446766027577162,
+      "learning_rate": 2.3107438016528924e-07,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -444.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0106,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.0302734375,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -11.25,
+      "step": 1746
+    },
+    {
+      "epoch": 2.143558282208589,
+      "grad_norm": 3.46516331869763,
+      "learning_rate": 2.3074380165289256e-07,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1826171875,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.6875,
+      "step": 1747
+    },
+    {
+      "epoch": 2.1447852760736197,
+      "grad_norm": 2.876651792706391,
+      "learning_rate": 2.3041322314049587e-07,
+      "logits/chosen": -0.35546875,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0085,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.43359375,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -11.625,
+      "step": 1748
+    },
+    {
+      "epoch": 2.1460122699386504,
+      "grad_norm": 3.5223225418119357,
+      "learning_rate": 2.3008264462809914e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0063,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.33203125,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -11.0625,
+      "step": 1749
+    },
+    {
+      "epoch": 2.147239263803681,
+      "grad_norm": 2.4641016521985453,
+      "learning_rate": 2.2975206611570246e-07,
+      "logits/chosen": -0.353515625,
+      "logits/rejected": -0.57421875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.25,
+      "step": 1750
+    },
+    {
+      "epoch": 2.1484662576687117,
+      "grad_norm": 2.201181311782617,
+      "learning_rate": 2.2942148760330577e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -322.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.349609375,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.9375,
+      "step": 1751
+    },
+    {
+      "epoch": 2.1496932515337424,
+      "grad_norm": 4.52660612383527,
+      "learning_rate": 2.290909090909091e-07,
+      "logits/chosen": -0.248046875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -456.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0073,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.330078125,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.0625,
+      "step": 1752
+    },
+    {
+      "epoch": 2.150920245398773,
+      "grad_norm": 2.7084686481091227,
+      "learning_rate": 2.287603305785124e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.236328125,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -11.1875,
+      "step": 1753
+    },
+    {
+      "epoch": 2.1521472392638037,
+      "grad_norm": 4.7459030428657805,
+      "learning_rate": 2.2842975206611567e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0105,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.625,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -11.8125,
+      "step": 1754
+    },
+    {
+      "epoch": 2.1533742331288344,
+      "grad_norm": 9.905820663459277,
+      "learning_rate": 2.28099173553719e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0147,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.049560546875,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.5,
+      "step": 1755
+    },
+    {
+      "epoch": 2.154601226993865,
+      "grad_norm": 3.539962554608615,
+      "learning_rate": 2.277685950413223e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0078,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.09912109375,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -11.5,
+      "step": 1756
+    },
+    {
+      "epoch": 2.1558282208588957,
+      "grad_norm": 2.638015802633454,
+      "learning_rate": 2.2743801652892563e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.119140625,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -11.0,
+      "step": 1757
+    },
+    {
+      "epoch": 2.1570552147239264,
+      "grad_norm": 2.1243833504173577,
+      "learning_rate": 2.271074380165289e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -332.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.34375,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -11.5,
+      "step": 1758
+    },
+    {
+      "epoch": 2.158282208588957,
+      "grad_norm": 3.380202144495586,
+      "learning_rate": 2.267768595041322e-07,
+      "logits/chosen": -0.435546875,
+      "logits/rejected": -0.6171875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.06201171875,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.1875,
+      "step": 1759
+    },
+    {
+      "epoch": 2.1595092024539877,
+      "grad_norm": 1.6297400292641515,
+      "learning_rate": 2.2644628099173553e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.54296875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -11.6875,
+      "step": 1760
+    },
+    {
+      "epoch": 2.1607361963190184,
+      "grad_norm": 15.921644122365354,
+      "learning_rate": 2.2611570247933885e-07,
+      "logits/chosen": -0.322265625,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0144,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.11279296875,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.0,
+      "step": 1761
+    },
+    {
+      "epoch": 2.161963190184049,
+      "grad_norm": 2.4655209270459673,
+      "learning_rate": 2.2578512396694216e-07,
+      "logits/chosen": -0.3984375,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0082,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.76953125,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -12.0,
+      "step": 1762
+    },
+    {
+      "epoch": 2.1631901840490797,
+      "grad_norm": 2.506079680679013,
+      "learning_rate": 2.2545454545454543e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0142,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.2265625,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -12.125,
+      "step": 1763
+    },
+    {
+      "epoch": 2.1644171779141104,
+      "grad_norm": 1.8737881956824924,
+      "learning_rate": 2.2512396694214875e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0218,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.337890625,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -11.125,
+      "step": 1764
+    },
+    {
+      "epoch": 2.165644171779141,
+      "grad_norm": 2.7050237131193113,
+      "learning_rate": 2.2479338842975206e-07,
+      "logits/chosen": -0.4453125,
+      "logits/rejected": -0.6640625,
+      "logps/chosen": -460.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0103,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.66015625,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.1875,
+      "step": 1765
+    },
+    {
+      "epoch": 2.1668711656441717,
+      "grad_norm": 2.257640770661483,
+      "learning_rate": 2.2446280991735538e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1201171875,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -11.0625,
+      "step": 1766
+    },
+    {
+      "epoch": 2.1680981595092024,
+      "grad_norm": 2.484936623159917,
+      "learning_rate": 2.2413223140495865e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0208740234375,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -11.875,
+      "step": 1767
+    },
+    {
+      "epoch": 2.169325153374233,
+      "grad_norm": 2.7281260389042323,
+      "learning_rate": 2.2380165289256196e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2353515625,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -11.5,
+      "step": 1768
+    },
+    {
+      "epoch": 2.1705521472392637,
+      "grad_norm": 4.139564056384544,
+      "learning_rate": 2.2347107438016528e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0065,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.44921875,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.5,
+      "step": 1769
+    },
+    {
+      "epoch": 2.1717791411042944,
+      "grad_norm": 4.640211167576865,
+      "learning_rate": 2.231404958677686e-07,
+      "logits/chosen": -0.47265625,
+      "logits/rejected": -0.62109375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0069,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.52734375,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -12.125,
+      "step": 1770
+    },
+    {
+      "epoch": 2.173006134969325,
+      "grad_norm": 2.250598004237111,
+      "learning_rate": 2.2280991735537192e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.412109375,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.3125,
+      "step": 1771
+    },
+    {
+      "epoch": 2.1742331288343557,
+      "grad_norm": 11.251999756810362,
+      "learning_rate": 2.2247933884297518e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0128,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.14453125,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -11.875,
+      "step": 1772
+    },
+    {
+      "epoch": 2.1754601226993864,
+      "grad_norm": 1.4276725163930015,
+      "learning_rate": 2.221487603305785e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.33203125,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -11.9375,
+      "step": 1773
+    },
+    {
+      "epoch": 2.176687116564417,
+      "grad_norm": 2.1180603938002274,
+      "learning_rate": 2.2181818181818182e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0091,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.080078125,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -11.5,
+      "step": 1774
+    },
+    {
+      "epoch": 2.1779141104294477,
+      "grad_norm": 1.6015084412639278,
+      "learning_rate": 2.2148760330578513e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.337890625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.9375,
+      "step": 1775
+    },
+    {
+      "epoch": 2.1791411042944784,
+      "grad_norm": 4.266753722558952,
+      "learning_rate": 2.211570247933884e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.09228515625,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.375,
+      "step": 1776
+    },
+    {
+      "epoch": 2.180368098159509,
+      "grad_norm": 1.404346018065358,
+      "learning_rate": 2.2082644628099172e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2294921875,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.1875,
+      "step": 1777
+    },
+    {
+      "epoch": 2.1815950920245397,
+      "grad_norm": 1.6848584885630067,
+      "learning_rate": 2.2049586776859503e-07,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0101,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.25390625,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.0625,
+      "step": 1778
+    },
+    {
+      "epoch": 2.1828220858895704,
+      "grad_norm": 1.4478353271672881,
+      "learning_rate": 2.2016528925619835e-07,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -446.0,
+      "logps/rejected": -404.0,
+      "loss": 0.005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.57421875,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.25,
+      "step": 1779
+    },
+    {
+      "epoch": 2.184049079754601,
+      "grad_norm": 2.0280244781674672,
+      "learning_rate": 2.1983471074380162e-07,
+      "logits/chosen": -0.40234375,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -432.0,
+      "loss": 0.0112,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.0703125,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.75,
+      "step": 1780
+    },
+    {
+      "epoch": 2.1852760736196317,
+      "grad_norm": 1.3856385755403737,
+      "learning_rate": 2.1950413223140493e-07,
+      "logits/chosen": -0.380859375,
+      "logits/rejected": -0.62890625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0084,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.37109375,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -12.875,
+      "step": 1781
+    },
+    {
+      "epoch": 2.1865030674846624,
+      "grad_norm": 3.1275442960896287,
+      "learning_rate": 2.1917355371900825e-07,
+      "logits/chosen": -0.36328125,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0079,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.51171875,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.375,
+      "step": 1782
+    },
+    {
+      "epoch": 2.187730061349693,
+      "grad_norm": 2.1395070210252727,
+      "learning_rate": 2.1884297520661157e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.31640625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.875,
+      "step": 1783
+    },
+    {
+      "epoch": 2.1889570552147237,
+      "grad_norm": 3.2612835541859155,
+      "learning_rate": 2.185123966942149e-07,
+      "logits/chosen": -0.466796875,
+      "logits/rejected": -0.69140625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0098,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.00885009765625,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.8125,
+      "step": 1784
+    },
+    {
+      "epoch": 2.190184049079755,
+      "grad_norm": 2.354013508970521,
+      "learning_rate": 2.1818181818181815e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.58984375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.75,
+      "step": 1785
+    },
+    {
+      "epoch": 2.191411042944785,
+      "grad_norm": 2.0889518399102025,
+      "learning_rate": 2.1785123966942147e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.349609375,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -11.5625,
+      "step": 1786
+    },
+    {
+      "epoch": 2.192638036809816,
+      "grad_norm": 1.7999851881845579,
+      "learning_rate": 2.175206611570248e-07,
+      "logits/chosen": -0.34765625,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -450.0,
+      "loss": 0.0154,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.09716796875,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -11.75,
+      "step": 1787
+    },
+    {
+      "epoch": 2.1938650306748464,
+      "grad_norm": 2.439699548469828,
+      "learning_rate": 2.171900826446281e-07,
+      "logits/chosen": -0.10595703125,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.09521484375,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -11.75,
+      "step": 1788
+    },
+    {
+      "epoch": 2.1950920245398775,
+      "grad_norm": 1.7086298027404399,
+      "learning_rate": 2.1685950413223137e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8203125,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -12.375,
+      "step": 1789
+    },
+    {
+      "epoch": 2.196319018404908,
+      "grad_norm": 4.723423388102118,
+      "learning_rate": 2.165289256198347e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0148,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.431640625,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -11.1875,
+      "step": 1790
+    },
+    {
+      "epoch": 2.197546012269939,
+      "grad_norm": 2.76556691818057,
+      "learning_rate": 2.16198347107438e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.76171875,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -12.0625,
+      "step": 1791
+    },
+    {
+      "epoch": 2.1987730061349695,
+      "grad_norm": 3.0843341292427593,
+      "learning_rate": 2.1586776859504132e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0089,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.97265625,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -12.0,
+      "step": 1792
+    },
+    {
+      "epoch": 2.2,
+      "grad_norm": 1.1939670310758894,
+      "learning_rate": 2.1553719008264464e-07,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.48046875,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -11.875,
+      "step": 1793
+    },
+    {
+      "epoch": 2.201226993865031,
+      "grad_norm": 1.1285543366658424,
+      "learning_rate": 2.152066115702479e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.19140625,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.375,
+      "step": 1794
+    },
+    {
+      "epoch": 2.2024539877300615,
+      "grad_norm": 1.3076871745986365,
+      "learning_rate": 2.1487603305785122e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -404.0,
+      "loss": 0.009,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.041259765625,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -11.3125,
+      "step": 1795
+    },
+    {
+      "epoch": 2.203680981595092,
+      "grad_norm": 2.7867554217261366,
+      "learning_rate": 2.1454545454545454e-07,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0078,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5859375,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -11.4375,
+      "step": 1796
+    },
+    {
+      "epoch": 2.204907975460123,
+      "grad_norm": 1.2738124654909748,
+      "learning_rate": 2.1421487603305786e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -398.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.265625,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -11.8125,
+      "step": 1797
+    },
+    {
+      "epoch": 2.2061349693251535,
+      "grad_norm": 2.098940233317881,
+      "learning_rate": 2.1388429752066112e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.396484375,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.4375,
+      "step": 1798
+    },
+    {
+      "epoch": 2.207361963190184,
+      "grad_norm": 2.809072826499818,
+      "learning_rate": 2.1355371900826444e-07,
+      "logits/chosen": -0.24609375,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0073,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1611328125,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.1875,
+      "step": 1799
+    },
+    {
+      "epoch": 2.208588957055215,
+      "grad_norm": 2.148612977924567,
+      "learning_rate": 2.1322314049586776e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0058,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.55078125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.8125,
+      "step": 1800
+    },
+    {
+      "epoch": 2.2098159509202455,
+      "grad_norm": 3.278545376467699,
+      "learning_rate": 2.1289256198347108e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0072,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.302734375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.5,
+      "step": 1801
+    },
+    {
+      "epoch": 2.211042944785276,
+      "grad_norm": 1.848338028344327,
+      "learning_rate": 2.125619834710744e-07,
+      "logits/chosen": -0.376953125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2138671875,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -11.5625,
+      "step": 1802
+    },
+    {
+      "epoch": 2.212269938650307,
+      "grad_norm": 1.7308475472556093,
+      "learning_rate": 2.1223140495867766e-07,
+      "logits/chosen": -0.423828125,
+      "logits/rejected": -0.64453125,
+      "logps/chosen": -398.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.048583984375,
+      "rewards/margins": 13.1875,
+      "rewards/rejected": -13.125,
+      "step": 1803
+    },
+    {
+      "epoch": 2.2134969325153375,
+      "grad_norm": 1.3449939370670436,
+      "learning_rate": 2.1190082644628098e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.10205078125,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.5625,
+      "step": 1804
+    },
+    {
+      "epoch": 2.214723926380368,
+      "grad_norm": 2.3383041071303574,
+      "learning_rate": 2.115702479338843e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -448.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1474609375,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.3125,
+      "step": 1805
+    },
+    {
+      "epoch": 2.215950920245399,
+      "grad_norm": 2.665963413461305,
+      "learning_rate": 2.112396694214876e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0058,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.96484375,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.6875,
+      "step": 1806
+    },
+    {
+      "epoch": 2.2171779141104295,
+      "grad_norm": 2.913294776622937,
+      "learning_rate": 2.1090909090909088e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0068,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1513671875,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.4375,
+      "step": 1807
+    },
+    {
+      "epoch": 2.21840490797546,
+      "grad_norm": 1.245001301546518,
+      "learning_rate": 2.105785123966942e-07,
+      "logits/chosen": -0.390625,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -418.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.030517578125,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -12.625,
+      "step": 1808
+    },
+    {
+      "epoch": 2.219631901840491,
+      "grad_norm": 1.6750743607205356,
+      "learning_rate": 2.102479338842975e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -428.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.515625,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.5,
+      "step": 1809
+    },
+    {
+      "epoch": 2.2208588957055215,
+      "grad_norm": 1.4805636986275978,
+      "learning_rate": 2.0991735537190083e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0008544921875,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -11.375,
+      "step": 1810
+    },
+    {
+      "epoch": 2.222085889570552,
+      "grad_norm": 2.461416050881252,
+      "learning_rate": 2.0958677685950415e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0062,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.984375,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -12.1875,
+      "step": 1811
+    },
+    {
+      "epoch": 2.223312883435583,
+      "grad_norm": 1.7174057427207048,
+      "learning_rate": 2.092561983471074e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -370.0,
+      "loss": 0.014,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.048583984375,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.1875,
+      "step": 1812
+    },
+    {
+      "epoch": 2.2245398773006135,
+      "grad_norm": 3.3203073185053587,
+      "learning_rate": 2.0892561983471073e-07,
+      "logits/chosen": -0.404296875,
+      "logits/rejected": -0.61328125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2578125,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -11.75,
+      "step": 1813
+    },
+    {
+      "epoch": 2.225766871165644,
+      "grad_norm": 1.816441340655354,
+      "learning_rate": 2.0859504132231405e-07,
+      "logits/chosen": -0.35546875,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0121,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.025146484375,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -11.6875,
+      "step": 1814
+    },
+    {
+      "epoch": 2.226993865030675,
+      "grad_norm": 2.598389848167637,
+      "learning_rate": 2.0826446280991737e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1015625,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.3125,
+      "step": 1815
+    },
+    {
+      "epoch": 2.2282208588957055,
+      "grad_norm": 2.8347932585069384,
+      "learning_rate": 2.0793388429752063e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0125,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.15234375,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -11.4375,
+      "step": 1816
+    },
+    {
+      "epoch": 2.229447852760736,
+      "grad_norm": 1.3924644368445442,
+      "learning_rate": 2.0760330578512395e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.5,
+      "logps/chosen": -440.0,
+      "logps/rejected": -398.0,
+      "loss": 0.009,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.326171875,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -11.875,
+      "step": 1817
+    },
+    {
+      "epoch": 2.230674846625767,
+      "grad_norm": 2.8078751744272874,
+      "learning_rate": 2.0727272727272727e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0118,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.328125,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.125,
+      "step": 1818
+    },
+    {
+      "epoch": 2.2319018404907975,
+      "grad_norm": 3.139070091749575,
+      "learning_rate": 2.0694214876033058e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0077,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.82421875,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -11.5625,
+      "step": 1819
+    },
+    {
+      "epoch": 2.233128834355828,
+      "grad_norm": 1.4470298427585824,
+      "learning_rate": 2.0661157024793385e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -340.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.059326171875,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -10.8125,
+      "step": 1820
+    },
+    {
+      "epoch": 2.234355828220859,
+      "grad_norm": 1.9049458977820766,
+      "learning_rate": 2.0628099173553717e-07,
+      "logits/chosen": -0.369140625,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.019287109375,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.125,
+      "step": 1821
+    },
+    {
+      "epoch": 2.2355828220858895,
+      "grad_norm": 2.4336548090189276,
+      "learning_rate": 2.0595041322314048e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.026123046875,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -11.4375,
+      "step": 1822
+    },
+    {
+      "epoch": 2.23680981595092,
+      "grad_norm": 2.444636911980317,
+      "learning_rate": 2.056198347107438e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -450.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0072,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.478515625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.125,
+      "step": 1823
+    },
+    {
+      "epoch": 2.238036809815951,
+      "grad_norm": 1.2139521824128825,
+      "learning_rate": 2.0528925619834712e-07,
+      "logits/chosen": -0.10888671875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.11572265625,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -11.1875,
+      "step": 1824
+    },
+    {
+      "epoch": 2.2392638036809815,
+      "grad_norm": 2.832909044059329,
+      "learning_rate": 2.0495867768595038e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.57421875,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -11.875,
+      "step": 1825
+    },
+    {
+      "epoch": 2.240490797546012,
+      "grad_norm": 1.9208108567871578,
+      "learning_rate": 2.046280991735537e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.70703125,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -12.0,
+      "step": 1826
+    },
+    {
+      "epoch": 2.241717791411043,
+      "grad_norm": 1.7817245629392895,
+      "learning_rate": 2.0429752066115702e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.052978515625,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.25,
+      "step": 1827
+    },
+    {
+      "epoch": 2.2429447852760735,
+      "grad_norm": 2.1872037368082995,
+      "learning_rate": 2.0396694214876034e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.058837890625,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -13.0,
+      "step": 1828
+    },
+    {
+      "epoch": 2.244171779141104,
+      "grad_norm": 2.8211755197465127,
+      "learning_rate": 2.036363636363636e-07,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0069,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.62890625,
+      "rewards/margins": 10.75,
+      "rewards/rejected": -11.4375,
+      "step": 1829
+    },
+    {
+      "epoch": 2.245398773006135,
+      "grad_norm": 4.221632653865993,
+      "learning_rate": 2.0330578512396692e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0086,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.412109375,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -11.8125,
+      "step": 1830
+    },
+    {
+      "epoch": 2.2466257668711656,
+      "grad_norm": 1.176306021065417,
+      "learning_rate": 2.0297520661157024e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1953125,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.9375,
+      "step": 1831
+    },
+    {
+      "epoch": 2.247852760736196,
+      "grad_norm": 2.451644772089647,
+      "learning_rate": 2.0264462809917356e-07,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -488.0,
+      "logps/rejected": -444.0,
+      "loss": 0.0068,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1591796875,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.25,
+      "step": 1832
+    },
+    {
+      "epoch": 2.249079754601227,
+      "grad_norm": 1.6116570146673836,
+      "learning_rate": 2.0231404958677687e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0094,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.06884765625,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -12.75,
+      "step": 1833
+    },
+    {
+      "epoch": 2.2503067484662576,
+      "grad_norm": 1.0635505638936062,
+      "learning_rate": 2.0198347107438014e-07,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.54296875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -384.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.33984375,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.9375,
+      "step": 1834
+    },
+    {
+      "epoch": 2.2515337423312882,
+      "grad_norm": 0.9893953082570133,
+      "learning_rate": 2.0165289256198346e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -394.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.275390625,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.0625,
+      "step": 1835
+    },
+    {
+      "epoch": 2.252760736196319,
+      "grad_norm": 1.0250426120552636,
+      "learning_rate": 2.0132231404958677e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -342.0,
+      "logps/rejected": -336.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.08544921875,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -11.9375,
+      "step": 1836
+    },
+    {
+      "epoch": 2.2539877300613496,
+      "grad_norm": 2.8278181518926564,
+      "learning_rate": 2.009917355371901e-07,
+      "logits/chosen": -0.37109375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.70703125,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.6875,
+      "step": 1837
+    },
+    {
+      "epoch": 2.2552147239263802,
+      "grad_norm": 1.4733479526766848,
+      "learning_rate": 2.0066115702479336e-07,
+      "logits/chosen": -0.345703125,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.212890625,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.0625,
+      "step": 1838
+    },
+    {
+      "epoch": 2.256441717791411,
+      "grad_norm": 3.4361238388087845,
+      "learning_rate": 2.0033057851239667e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0076,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.82421875,
+      "rewards/margins": 10.6875,
+      "rewards/rejected": -11.5,
+      "step": 1839
+    },
+    {
+      "epoch": 2.2576687116564416,
+      "grad_norm": 2.4736843264341632,
+      "learning_rate": 2e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0075,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1923828125,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -11.5625,
+      "step": 1840
+    },
+    {
+      "epoch": 2.2588957055214722,
+      "grad_norm": 8.48772894428233,
+      "learning_rate": 1.996694214876033e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0102,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -11.625,
+      "step": 1841
+    },
+    {
+      "epoch": 2.260122699386503,
+      "grad_norm": 1.566468300969243,
+      "learning_rate": 1.993388429752066e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -396.0,
+      "loss": 0.01,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.7109375,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.9375,
+      "step": 1842
+    },
+    {
+      "epoch": 2.2613496932515336,
+      "grad_norm": 6.181673023254863,
+      "learning_rate": 1.9900826446280992e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0141,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.62109375,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -12.0625,
+      "step": 1843
+    },
+    {
+      "epoch": 2.2625766871165642,
+      "grad_norm": 1.8486368399690625,
+      "learning_rate": 1.986776859504132e-07,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -440.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.07958984375,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.5,
+      "step": 1844
+    },
+    {
+      "epoch": 2.263803680981595,
+      "grad_norm": 1.733658432437379,
+      "learning_rate": 1.9834710743801653e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -436.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -11.25,
+      "step": 1845
+    },
+    {
+      "epoch": 2.265030674846626,
+      "grad_norm": 2.5246169361664146,
+      "learning_rate": 1.9801652892561982e-07,
+      "logits/chosen": -0.353515625,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -446.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0056,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.28125,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.1875,
+      "step": 1846
+    },
+    {
+      "epoch": 2.2662576687116562,
+      "grad_norm": 1.4894831664747705,
+      "learning_rate": 1.9768595041322314e-07,
+      "logits/chosen": -0.490234375,
+      "logits/rejected": -0.62109375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -434.0,
+      "loss": 0.009,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.490234375,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.25,
+      "step": 1847
+    },
+    {
+      "epoch": 2.2674846625766873,
+      "grad_norm": 2.0207780327723377,
+      "learning_rate": 1.9735537190082643e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0091,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.9296875,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -10.875,
+      "step": 1848
+    },
+    {
+      "epoch": 2.2687116564417176,
+      "grad_norm": 1.2258627826108883,
+      "learning_rate": 1.9702479338842974e-07,
+      "logits/chosen": -0.365234375,
+      "logits/rejected": -0.609375,
+      "logps/chosen": -464.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.76953125,
+      "rewards/margins": 13.4375,
+      "rewards/rejected": -12.6875,
+      "step": 1849
+    },
+    {
+      "epoch": 2.2699386503067487,
+      "grad_norm": 3.0800268989197837,
+      "learning_rate": 1.9669421487603306e-07,
+      "logits/chosen": -0.35546875,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0142,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.330078125,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -11.4375,
+      "step": 1850
+    },
+    {
+      "epoch": 2.271165644171779,
+      "grad_norm": 1.887638172215392,
+      "learning_rate": 1.9636363636363635e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.126953125,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.25,
+      "step": 1851
+    },
+    {
+      "epoch": 2.27239263803681,
+      "grad_norm": 2.913188428112757,
+      "learning_rate": 1.9603305785123967e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -452.0,
+      "logps/rejected": -440.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.263671875,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.1875,
+      "step": 1852
+    },
+    {
+      "epoch": 2.2736196319018402,
+      "grad_norm": 3.519004572779955,
+      "learning_rate": 1.9570247933884296e-07,
+      "logits/chosen": -0.0869140625,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0088,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6484375,
+      "rewards/margins": 11.125,
+      "rewards/rejected": -11.8125,
+      "step": 1853
+    },
+    {
+      "epoch": 2.2748466257668714,
+      "grad_norm": 2.6955681207287205,
+      "learning_rate": 1.9537190082644628e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0057,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.30859375,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -11.625,
+      "step": 1854
+    },
+    {
+      "epoch": 2.276073619631902,
+      "grad_norm": 2.200933623924347,
+      "learning_rate": 1.9504132231404957e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -444.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.30078125,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.625,
+      "step": 1855
+    },
+    {
+      "epoch": 2.2773006134969327,
+      "grad_norm": 2.6992394253423906,
+      "learning_rate": 1.947107438016529e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2080078125,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.125,
+      "step": 1856
+    },
+    {
+      "epoch": 2.2785276073619634,
+      "grad_norm": 1.527055694334907,
+      "learning_rate": 1.9438016528925618e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0038,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -12.1875,
+      "step": 1857
+    },
+    {
+      "epoch": 2.279754601226994,
+      "grad_norm": 2.5699695578569357,
+      "learning_rate": 1.940495867768595e-07,
+      "logits/chosen": -0.3984375,
+      "logits/rejected": -0.54296875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0098,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.48828125,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.0625,
+      "step": 1858
+    },
+    {
+      "epoch": 2.2809815950920247,
+      "grad_norm": 1.652532936070525,
+      "learning_rate": 1.937190082644628e-07,
+      "logits/chosen": -0.4609375,
+      "logits/rejected": -0.609375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -398.0,
+      "loss": 0.005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.0703125,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.25,
+      "step": 1859
+    },
+    {
+      "epoch": 2.2822085889570554,
+      "grad_norm": 2.046865864987721,
+      "learning_rate": 1.933884297520661e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.5546875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.10986328125,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -11.75,
+      "step": 1860
+    },
+    {
+      "epoch": 2.283435582822086,
+      "grad_norm": 4.718724960720358,
+      "learning_rate": 1.9305785123966942e-07,
+      "logits/chosen": -0.40625,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0114,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.62890625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -12.1875,
+      "step": 1861
+    },
+    {
+      "epoch": 2.2846625766871167,
+      "grad_norm": 3.272853368590046,
+      "learning_rate": 1.9272727272727272e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0061,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2890625,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -11.75,
+      "step": 1862
+    },
+    {
+      "epoch": 2.2858895705521474,
+      "grad_norm": 2.8764124765662,
+      "learning_rate": 1.9239669421487603e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -436.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1025390625,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.125,
+      "step": 1863
+    },
+    {
+      "epoch": 2.287116564417178,
+      "grad_norm": 1.4543643006828244,
+      "learning_rate": 1.9206611570247932e-07,
+      "logits/chosen": -0.357421875,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -440.0,
+      "loss": 0.0089,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.470703125,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.25,
+      "step": 1864
+    },
+    {
+      "epoch": 2.2883435582822087,
+      "grad_norm": 1.317051212251163,
+      "learning_rate": 1.9173553719008264e-07,
+      "logits/chosen": -0.36328125,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -410.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1494140625,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.1875,
+      "step": 1865
+    },
+    {
+      "epoch": 2.2895705521472394,
+      "grad_norm": 4.629396065761951,
+      "learning_rate": 1.9140495867768593e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0079,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.84765625,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.25,
+      "step": 1866
+    },
+    {
+      "epoch": 2.29079754601227,
+      "grad_norm": 1.1681464908141868,
+      "learning_rate": 1.9107438016528925e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0087,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.08251953125,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.8125,
+      "step": 1867
+    },
+    {
+      "epoch": 2.2920245398773007,
+      "grad_norm": 1.2828685568956375,
+      "learning_rate": 1.9074380165289254e-07,
+      "logits/chosen": -0.35546875,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6328125,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.375,
+      "step": 1868
+    },
+    {
+      "epoch": 2.2932515337423314,
+      "grad_norm": 0.7850707353580862,
+      "learning_rate": 1.9041322314049586e-07,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -420.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0028,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.41015625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -12.0,
+      "step": 1869
+    },
+    {
+      "epoch": 2.294478527607362,
+      "grad_norm": 1.417187394797712,
+      "learning_rate": 1.9008264462809918e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.494140625,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.5,
+      "step": 1870
+    },
+    {
+      "epoch": 2.2957055214723927,
+      "grad_norm": 2.354389956684456,
+      "learning_rate": 1.8975206611570247e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5703125,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.625,
+      "step": 1871
+    },
+    {
+      "epoch": 2.2969325153374234,
+      "grad_norm": 1.532873147921412,
+      "learning_rate": 1.894214876033058e-07,
+      "logits/chosen": -0.392578125,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.26171875,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.3125,
+      "step": 1872
+    },
+    {
+      "epoch": 2.298159509202454,
+      "grad_norm": 1.8213986312542025,
+      "learning_rate": 1.8909090909090908e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.921875,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -11.8125,
+      "step": 1873
+    },
+    {
+      "epoch": 2.2993865030674847,
+      "grad_norm": 1.9588542117678927,
+      "learning_rate": 1.887603305785124e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -446.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.490234375,
+      "rewards/margins": 13.75,
+      "rewards/rejected": -13.25,
+      "step": 1874
+    },
+    {
+      "epoch": 2.3006134969325154,
+      "grad_norm": 1.639111133136382,
+      "learning_rate": 1.884297520661157e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.99609375,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.75,
+      "step": 1875
+    },
+    {
+      "epoch": 2.301840490797546,
+      "grad_norm": 5.211392012726367,
+      "learning_rate": 1.88099173553719e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -466.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0092,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.12109375,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.6875,
+      "step": 1876
+    },
+    {
+      "epoch": 2.3030674846625767,
+      "grad_norm": 2.621493088958013,
+      "learning_rate": 1.877685950413223e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0057,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0986328125,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -10.9375,
+      "step": 1877
+    },
+    {
+      "epoch": 2.3042944785276074,
+      "grad_norm": 2.792412143684038,
+      "learning_rate": 1.8743801652892561e-07,
+      "logits/chosen": -0.353515625,
+      "logits/rejected": -0.60546875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -398.0,
+      "loss": 0.006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.08447265625,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -11.8125,
+      "step": 1878
+    },
+    {
+      "epoch": 2.305521472392638,
+      "grad_norm": 2.3872448294497604,
+      "learning_rate": 1.871074380165289e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -396.0,
+      "loss": 0.01,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.46484375,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.1875,
+      "step": 1879
+    },
+    {
+      "epoch": 2.3067484662576687,
+      "grad_norm": 2.563689648421724,
+      "learning_rate": 1.8677685950413222e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -11.5625,
+      "step": 1880
+    },
+    {
+      "epoch": 2.3079754601226994,
+      "grad_norm": 2.4204693331054745,
+      "learning_rate": 1.8644628099173554e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1865234375,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -11.1875,
+      "step": 1881
+    },
+    {
+      "epoch": 2.30920245398773,
+      "grad_norm": 2.0606685117293404,
+      "learning_rate": 1.8611570247933883e-07,
+      "logits/chosen": -0.361328125,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0104,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.275390625,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -11.875,
+      "step": 1882
+    },
+    {
+      "epoch": 2.3104294478527607,
+      "grad_norm": 2.4127012019087077,
+      "learning_rate": 1.8578512396694215e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -360.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0068,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.455078125,
+      "rewards/margins": 10.4375,
+      "rewards/rejected": -10.875,
+      "step": 1883
+    },
+    {
+      "epoch": 2.3116564417177914,
+      "grad_norm": 2.0779963798448007,
+      "learning_rate": 1.8545454545454544e-07,
+      "logits/chosen": -0.2001953125,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1083984375,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -11.625,
+      "step": 1884
+    },
+    {
+      "epoch": 2.312883435582822,
+      "grad_norm": 8.910863711571507,
+      "learning_rate": 1.8512396694214876e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -396.0,
+      "loss": 0.013,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.41796875,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.5625,
+      "step": 1885
+    },
+    {
+      "epoch": 2.3141104294478527,
+      "grad_norm": 2.157930567131376,
+      "learning_rate": 1.8479338842975205e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.375,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.25,
+      "step": 1886
+    },
+    {
+      "epoch": 2.3153374233128834,
+      "grad_norm": 1.998581678846594,
+      "learning_rate": 1.8446280991735537e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -448.0,
+      "logps/rejected": -462.0,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1923828125,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.375,
+      "step": 1887
+    },
+    {
+      "epoch": 2.316564417177914,
+      "grad_norm": 1.8309347062681074,
+      "learning_rate": 1.8413223140495866e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -434.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.37109375,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.0,
+      "step": 1888
+    },
+    {
+      "epoch": 2.3177914110429447,
+      "grad_norm": 1.5494876315219375,
+      "learning_rate": 1.8380165289256198e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.298828125,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.3125,
+      "step": 1889
+    },
+    {
+      "epoch": 2.3190184049079754,
+      "grad_norm": 1.8004590334466806,
+      "learning_rate": 1.834710743801653e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.54296875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.04638671875,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.0,
+      "step": 1890
+    },
+    {
+      "epoch": 2.320245398773006,
+      "grad_norm": 4.578360727254546,
+      "learning_rate": 1.8314049586776858e-07,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0077,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.40625,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -12.75,
+      "step": 1891
+    },
+    {
+      "epoch": 2.3214723926380367,
+      "grad_norm": 3.1319049802975334,
+      "learning_rate": 1.828099173553719e-07,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -430.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0085,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1689453125,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.9375,
+      "step": 1892
+    },
+    {
+      "epoch": 2.3226993865030674,
+      "grad_norm": 4.32793831862069,
+      "learning_rate": 1.824793388429752e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0078,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2470703125,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.125,
+      "step": 1893
+    },
+    {
+      "epoch": 2.323926380368098,
+      "grad_norm": 3.0787233114593398,
+      "learning_rate": 1.821487603305785e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0058,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.50390625,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -12.0,
+      "step": 1894
+    },
+    {
+      "epoch": 2.3251533742331287,
+      "grad_norm": 1.7015319499641888,
+      "learning_rate": 1.818181818181818e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.70703125,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.875,
+      "step": 1895
+    },
+    {
+      "epoch": 2.3263803680981594,
+      "grad_norm": 1.304284853691511,
+      "learning_rate": 1.8148760330578512e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0095,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.0167236328125,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.3125,
+      "step": 1896
+    },
+    {
+      "epoch": 2.32760736196319,
+      "grad_norm": 1.6399886275731839,
+      "learning_rate": 1.811570247933884e-07,
+      "logits/chosen": -0.40234375,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -404.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.10205078125,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -12.25,
+      "step": 1897
+    },
+    {
+      "epoch": 2.3288343558282207,
+      "grad_norm": 3.459045299533506,
+      "learning_rate": 1.8082644628099173e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0078,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.4453125,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -13.25,
+      "step": 1898
+    },
+    {
+      "epoch": 2.3300613496932514,
+      "grad_norm": 1.6396245671760088,
+      "learning_rate": 1.8049586776859502e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0097,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.640625,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -13.125,
+      "step": 1899
+    },
+    {
+      "epoch": 2.331288343558282,
+      "grad_norm": 2.808184212914399,
+      "learning_rate": 1.8016528925619834e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.392578125,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -11.75,
+      "step": 1900
+    },
+    {
+      "epoch": 2.3325153374233127,
+      "grad_norm": 1.5261987097468337,
+      "learning_rate": 1.7983471074380166e-07,
+      "logits/chosen": -0.427734375,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.38671875,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.9375,
+      "step": 1901
+    },
+    {
+      "epoch": 2.3337423312883434,
+      "grad_norm": 1.2099329432684087,
+      "learning_rate": 1.7950413223140495e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -446.0,
+      "logps/rejected": -442.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0361328125,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -11.75,
+      "step": 1902
+    },
+    {
+      "epoch": 2.334969325153374,
+      "grad_norm": 7.82004291553912,
+      "learning_rate": 1.7917355371900826e-07,
+      "logits/chosen": -0.330078125,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -432.0,
+      "loss": 0.0085,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.2392578125,
+      "rewards/margins": 13.625,
+      "rewards/rejected": -13.375,
+      "step": 1903
+    },
+    {
+      "epoch": 2.3361963190184047,
+      "grad_norm": 1.810917654076522,
+      "learning_rate": 1.7884297520661156e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.038818359375,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -11.9375,
+      "step": 1904
+    },
+    {
+      "epoch": 2.3374233128834354,
+      "grad_norm": 1.6784492814627991,
+      "learning_rate": 1.7851239669421487e-07,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.5,
+      "logps/chosen": -426.0,
+      "logps/rejected": -454.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1376953125,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -13.1875,
+      "step": 1905
+    },
+    {
+      "epoch": 2.338650306748466,
+      "grad_norm": 1.3884136999375363,
+      "learning_rate": 1.7818181818181817e-07,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1328125,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.875,
+      "step": 1906
+    },
+    {
+      "epoch": 2.3398773006134967,
+      "grad_norm": 2.0325371383269184,
+      "learning_rate": 1.7785123966942148e-07,
+      "logits/chosen": -0.431640625,
+      "logits/rejected": -0.65625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -410.0,
+      "loss": 0.006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.007476806640625,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.8125,
+      "step": 1907
+    },
+    {
+      "epoch": 2.3411042944785274,
+      "grad_norm": 1.339265017962584,
+      "learning_rate": 1.7752066115702477e-07,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.248046875,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -13.0625,
+      "step": 1908
+    },
+    {
+      "epoch": 2.3423312883435585,
+      "grad_norm": 1.2451318671903517,
+      "learning_rate": 1.771900826446281e-07,
+      "logits/chosen": -0.3828125,
+      "logits/rejected": -0.57421875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -12.0,
+      "step": 1909
+    },
+    {
+      "epoch": 2.3435582822085887,
+      "grad_norm": 1.5277783538430403,
+      "learning_rate": 1.768595041322314e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.13671875,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -11.875,
+      "step": 1910
+    },
+    {
+      "epoch": 2.34478527607362,
+      "grad_norm": 3.0517347841047022,
+      "learning_rate": 1.765289256198347e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0164,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.27734375,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.3125,
+      "step": 1911
+    },
+    {
+      "epoch": 2.34601226993865,
+      "grad_norm": 2.3677825785794595,
+      "learning_rate": 1.7619834710743802e-07,
+      "logits/chosen": -0.376953125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0061,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5390625,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -11.875,
+      "step": 1912
+    },
+    {
+      "epoch": 2.347239263803681,
+      "grad_norm": 1.5584135109112802,
+      "learning_rate": 1.758677685950413e-07,
+      "logits/chosen": -0.345703125,
+      "logits/rejected": -0.61328125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -362.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.41015625,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.25,
+      "step": 1913
+    },
+    {
+      "epoch": 2.3484662576687114,
+      "grad_norm": 2.362279418536187,
+      "learning_rate": 1.7553719008264463e-07,
+      "logits/chosen": -0.38671875,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2890625,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.25,
+      "step": 1914
+    },
+    {
+      "epoch": 2.3496932515337425,
+      "grad_norm": 3.1294817158158863,
+      "learning_rate": 1.7520661157024792e-07,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.439453125,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -12.875,
+      "step": 1915
+    },
+    {
+      "epoch": 2.3509202453987728,
+      "grad_norm": 1.18110352841037,
+      "learning_rate": 1.7487603305785124e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -438.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.220703125,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.625,
+      "step": 1916
+    },
+    {
+      "epoch": 2.352147239263804,
+      "grad_norm": 2.7639065395248648,
+      "learning_rate": 1.7454545454545453e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5390625,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.25,
+      "step": 1917
+    },
+    {
+      "epoch": 2.3533742331288345,
+      "grad_norm": 4.465262870874975,
+      "learning_rate": 1.7421487603305785e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.55078125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0104,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2314453125,
+      "rewards/margins": 13.625,
+      "rewards/rejected": -13.8125,
+      "step": 1918
+    },
+    {
+      "epoch": 2.354601226993865,
+      "grad_norm": 3.9312866613067974,
+      "learning_rate": 1.7388429752066114e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0132,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.7578125,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -12.375,
+      "step": 1919
+    },
+    {
+      "epoch": 2.355828220858896,
+      "grad_norm": 3.650701361648802,
+      "learning_rate": 1.7355371900826445e-07,
+      "logits/chosen": -0.396484375,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0088,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.32421875,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.3125,
+      "step": 1920
+    },
+    {
+      "epoch": 2.3570552147239265,
+      "grad_norm": 2.535124763317117,
+      "learning_rate": 1.7322314049586777e-07,
+      "logits/chosen": -0.322265625,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -404.0,
+      "loss": 0.005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.15234375,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.6875,
+      "step": 1921
+    },
+    {
+      "epoch": 2.358282208588957,
+      "grad_norm": 1.169161093609363,
+      "learning_rate": 1.7289256198347106e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.361328125,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -11.9375,
+      "step": 1922
+    },
+    {
+      "epoch": 2.359509202453988,
+      "grad_norm": 2.471684271604459,
+      "learning_rate": 1.7256198347107438e-07,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.67578125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1748046875,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.4375,
+      "step": 1923
+    },
+    {
+      "epoch": 2.3607361963190185,
+      "grad_norm": 1.5047808269029972,
+      "learning_rate": 1.7223140495867767e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.87890625,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -11.625,
+      "step": 1924
+    },
+    {
+      "epoch": 2.361963190184049,
+      "grad_norm": 1.8221698614181348,
+      "learning_rate": 1.71900826446281e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.34375,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -11.6875,
+      "step": 1925
+    },
+    {
+      "epoch": 2.36319018404908,
+      "grad_norm": 6.086203637613619,
+      "learning_rate": 1.7157024793388428e-07,
+      "logits/chosen": -0.40234375,
+      "logits/rejected": -0.6171875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.0087890625,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.75,
+      "step": 1926
+    },
+    {
+      "epoch": 2.3644171779141105,
+      "grad_norm": 1.987525129881384,
+      "learning_rate": 1.712396694214876e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.357421875,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.0625,
+      "step": 1927
+    },
+    {
+      "epoch": 2.365644171779141,
+      "grad_norm": 2.3514185268239185,
+      "learning_rate": 1.709090909090909e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0065,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.388671875,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -12.25,
+      "step": 1928
+    },
+    {
+      "epoch": 2.366871165644172,
+      "grad_norm": 2.3764735994154416,
+      "learning_rate": 1.705785123966942e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0109,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.65625,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.625,
+      "step": 1929
+    },
+    {
+      "epoch": 2.3680981595092025,
+      "grad_norm": 2.5270195773908033,
+      "learning_rate": 1.7024793388429753e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.58984375,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -12.1875,
+      "step": 1930
+    },
+    {
+      "epoch": 2.369325153374233,
+      "grad_norm": 1.2392383713400654,
+      "learning_rate": 1.6991735537190082e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.62890625,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.375,
+      "step": 1931
+    },
+    {
+      "epoch": 2.370552147239264,
+      "grad_norm": 2.286547983805737,
+      "learning_rate": 1.6958677685950413e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -438.0,
+      "loss": 0.006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.095703125,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -11.875,
+      "step": 1932
+    },
+    {
+      "epoch": 2.3717791411042946,
+      "grad_norm": 1.3837435251514467,
+      "learning_rate": 1.6925619834710743e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0888671875,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -10.6875,
+      "step": 1933
+    },
+    {
+      "epoch": 2.373006134969325,
+      "grad_norm": 3.301946066068422,
+      "learning_rate": 1.6892561983471074e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -442.0,
+      "logps/rejected": -448.0,
+      "loss": 0.006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.158203125,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -12.875,
+      "step": 1934
+    },
+    {
+      "epoch": 2.374233128834356,
+      "grad_norm": 1.2966282555983115,
+      "learning_rate": 1.6859504132231403e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2421875,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.0,
+      "step": 1935
+    },
+    {
+      "epoch": 2.3754601226993866,
+      "grad_norm": 1.0116513246009429,
+      "learning_rate": 1.6826446280991735e-07,
+      "logits/chosen": -0.345703125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -348.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0078,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.609375,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.5,
+      "step": 1936
+    },
+    {
+      "epoch": 2.3766871165644172,
+      "grad_norm": 2.0939092543887856,
+      "learning_rate": 1.6793388429752064e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.205078125,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.0,
+      "step": 1937
+    },
+    {
+      "epoch": 2.377914110429448,
+      "grad_norm": 2.9685778370718356,
+      "learning_rate": 1.6760330578512396e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -434.0,
+      "loss": 0.005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.23046875,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.6875,
+      "step": 1938
+    },
+    {
+      "epoch": 2.3791411042944786,
+      "grad_norm": 1.9992687759165277,
+      "learning_rate": 1.6727272727272725e-07,
+      "logits/chosen": -0.427734375,
+      "logits/rejected": -0.61328125,
+      "logps/chosen": -398.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.062255859375,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -13.125,
+      "step": 1939
+    },
+    {
+      "epoch": 2.3803680981595092,
+      "grad_norm": 1.3833417100595666,
+      "learning_rate": 1.6694214876033057e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -388.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.142578125,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -11.1875,
+      "step": 1940
+    },
+    {
+      "epoch": 2.38159509202454,
+      "grad_norm": 1.3216920973672437,
+      "learning_rate": 1.666115702479339e-07,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.0595703125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.3125,
+      "step": 1941
+    },
+    {
+      "epoch": 2.3828220858895706,
+      "grad_norm": 9.784985590846373,
+      "learning_rate": 1.6628099173553718e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0136,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.68359375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -11.5625,
+      "step": 1942
+    },
+    {
+      "epoch": 2.3840490797546012,
+      "grad_norm": 1.7249358716378238,
+      "learning_rate": 1.659504132231405e-07,
+      "logits/chosen": -0.36328125,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -450.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.9765625,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -12.125,
+      "step": 1943
+    },
+    {
+      "epoch": 2.385276073619632,
+      "grad_norm": 3.5603049517569927,
+      "learning_rate": 1.656198347107438e-07,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0072,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.55859375,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.375,
+      "step": 1944
+    },
+    {
+      "epoch": 2.3865030674846626,
+      "grad_norm": 5.963629046801992,
+      "learning_rate": 1.652892561983471e-07,
+      "logits/chosen": -0.359375,
+      "logits/rejected": -0.6171875,
+      "logps/chosen": -430.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0073,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -12.1875,
+      "step": 1945
+    },
+    {
+      "epoch": 2.3877300613496932,
+      "grad_norm": 4.848705261276623,
+      "learning_rate": 1.649586776859504e-07,
+      "logits/chosen": -0.09814453125,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -396.0,
+      "loss": 0.011,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.43359375,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -11.625,
+      "step": 1946
+    },
+    {
+      "epoch": 2.388957055214724,
+      "grad_norm": 0.8969608683768212,
+      "learning_rate": 1.6462809917355371e-07,
+      "logits/chosen": -0.375,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0022,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.212890625,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.5,
+      "step": 1947
+    },
+    {
+      "epoch": 2.3901840490797546,
+      "grad_norm": 3.1296435015020205,
+      "learning_rate": 1.64297520661157e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0058,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.197265625,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -11.6875,
+      "step": 1948
+    },
+    {
+      "epoch": 2.3914110429447852,
+      "grad_norm": 2.2915983533913975,
+      "learning_rate": 1.6396694214876032e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.203125,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -13.1875,
+      "step": 1949
+    },
+    {
+      "epoch": 2.392638036809816,
+      "grad_norm": 2.185559944792547,
+      "learning_rate": 1.6363636363636364e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -528.0,
+      "loss": 0.0082,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.361328125,
+      "rewards/margins": 24.625,
+      "rewards/rejected": -25.125,
+      "step": 1950
+    },
+    {
+      "epoch": 2.3938650306748466,
+      "grad_norm": 4.879741716801366,
+      "learning_rate": 1.6330578512396693e-07,
+      "logits/chosen": -0.09326171875,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0128,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.046875,
+      "rewards/margins": 9.3125,
+      "rewards/rejected": -10.375,
+      "step": 1951
+    },
+    {
+      "epoch": 2.3950920245398772,
+      "grad_norm": 2.9275259448191466,
+      "learning_rate": 1.6297520661157025e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -440.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0077,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.69140625,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.6875,
+      "step": 1952
+    },
+    {
+      "epoch": 2.396319018404908,
+      "grad_norm": 1.4413755433739257,
+      "learning_rate": 1.6264462809917354e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.57421875,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -12.9375,
+      "step": 1953
+    },
+    {
+      "epoch": 2.3975460122699386,
+      "grad_norm": 4.2391057595013795,
+      "learning_rate": 1.6231404958677686e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0064,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.189453125,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -12.4375,
+      "step": 1954
+    },
+    {
+      "epoch": 2.3987730061349692,
+      "grad_norm": 1.2302862638459167,
+      "learning_rate": 1.6198347107438015e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -452.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0083,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.150390625,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.1875,
+      "step": 1955
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 2.717664222533725,
+      "learning_rate": 1.6165289256198347e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -454.0,
+      "loss": 0.0058,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2236328125,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -12.875,
+      "step": 1956
+    },
+    {
+      "epoch": 2.4012269938650306,
+      "grad_norm": 2.216354779552163,
+      "learning_rate": 1.6132231404958676e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0064,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1669921875,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -11.8125,
+      "step": 1957
+    },
+    {
+      "epoch": 2.4024539877300612,
+      "grad_norm": 2.514057324524432,
+      "learning_rate": 1.6099173553719008e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.06591796875,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.0625,
+      "step": 1958
+    },
+    {
+      "epoch": 2.403680981595092,
+      "grad_norm": 1.258030788549785,
+      "learning_rate": 1.6066115702479337e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1484375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.375,
+      "step": 1959
+    },
+    {
+      "epoch": 2.4049079754601226,
+      "grad_norm": 1.9009758008954127,
+      "learning_rate": 1.6033057851239669e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0102,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.5703125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.875,
+      "step": 1960
+    },
+    {
+      "epoch": 2.4061349693251532,
+      "grad_norm": 1.977540043820977,
+      "learning_rate": 1.6e-07,
+      "logits/chosen": -0.3515625,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0104,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.134765625,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -13.25,
+      "step": 1961
+    },
+    {
+      "epoch": 2.407361963190184,
+      "grad_norm": 1.24091308995341,
+      "learning_rate": 1.596694214876033e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0028,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.373046875,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.8125,
+      "step": 1962
+    },
+    {
+      "epoch": 2.4085889570552146,
+      "grad_norm": 3.302779604929561,
+      "learning_rate": 1.593388429752066e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0057,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.8203125,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -11.6875,
+      "step": 1963
+    },
+    {
+      "epoch": 2.4098159509202453,
+      "grad_norm": 1.870659535045701,
+      "learning_rate": 1.590082644628099e-07,
+      "logits/chosen": -0.36328125,
+      "logits/rejected": -0.58203125,
+      "logps/chosen": -432.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.138671875,
+      "rewards/margins": 13.375,
+      "rewards/rejected": -13.25,
+      "step": 1964
+    },
+    {
+      "epoch": 2.411042944785276,
+      "grad_norm": 3.7334931302173593,
+      "learning_rate": 1.5867768595041322e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.12353515625,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.0625,
+      "step": 1965
+    },
+    {
+      "epoch": 2.4122699386503066,
+      "grad_norm": 0.9057345177874959,
+      "learning_rate": 1.583471074380165e-07,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.55078125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0022,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.29296875,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -12.3125,
+      "step": 1966
+    },
+    {
+      "epoch": 2.4134969325153373,
+      "grad_norm": 2.384986471776641,
+      "learning_rate": 1.5801652892561983e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0095,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.80859375,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.0625,
+      "step": 1967
+    },
+    {
+      "epoch": 2.414723926380368,
+      "grad_norm": 1.7432547417846278,
+      "learning_rate": 1.5768595041322312e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2119140625,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.4375,
+      "step": 1968
+    },
+    {
+      "epoch": 2.4159509202453986,
+      "grad_norm": 1.8150142109339031,
+      "learning_rate": 1.5735537190082644e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0038,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.060302734375,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -11.875,
+      "step": 1969
+    },
+    {
+      "epoch": 2.4171779141104293,
+      "grad_norm": 2.422415972988301,
+      "learning_rate": 1.5702479338842976e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0116,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.5234375,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.375,
+      "step": 1970
+    },
+    {
+      "epoch": 2.41840490797546,
+      "grad_norm": 4.462027721403495,
+      "learning_rate": 1.5669421487603305e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0069,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.625,
+      "rewards/margins": 11.125,
+      "rewards/rejected": -11.75,
+      "step": 1971
+    },
+    {
+      "epoch": 2.419631901840491,
+      "grad_norm": 2.7068110393521527,
+      "learning_rate": 1.5636363636363637e-07,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0101,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.19140625,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.625,
+      "step": 1972
+    },
+    {
+      "epoch": 2.4208588957055213,
+      "grad_norm": 0.9605198214469469,
+      "learning_rate": 1.5603305785123966e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0022,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.57421875,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.0,
+      "step": 1973
+    },
+    {
+      "epoch": 2.4220858895705524,
+      "grad_norm": 1.5080258091735799,
+      "learning_rate": 1.5570247933884297e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -446.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.10205078125,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -13.0625,
+      "step": 1974
+    },
+    {
+      "epoch": 2.4233128834355826,
+      "grad_norm": 3.790453778549161,
+      "learning_rate": 1.5537190082644627e-07,
+      "logits/chosen": -0.359375,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -440.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0064,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1767578125,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -12.4375,
+      "step": 1975
+    },
+    {
+      "epoch": 2.4245398773006137,
+      "grad_norm": 1.2072818871930608,
+      "learning_rate": 1.5504132231404958e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0091,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.59375,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.25,
+      "step": 1976
+    },
+    {
+      "epoch": 2.425766871165644,
+      "grad_norm": 2.7715355453038066,
+      "learning_rate": 1.5471074380165287e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.314453125,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.0625,
+      "step": 1977
+    },
+    {
+      "epoch": 2.426993865030675,
+      "grad_norm": 1.166908314086118,
+      "learning_rate": 1.543801652892562e-07,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -470.0,
+      "logps/rejected": -468.0,
+      "loss": 0.0024,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.67578125,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -13.6875,
+      "step": 1978
+    },
+    {
+      "epoch": 2.4282208588957057,
+      "grad_norm": 2.1162650911137337,
+      "learning_rate": 1.540495867768595e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -420.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0058,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5859375,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.9375,
+      "step": 1979
+    },
+    {
+      "epoch": 2.4294478527607364,
+      "grad_norm": 1.6599260550663577,
+      "learning_rate": 1.537190082644628e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.75,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.75,
+      "step": 1980
+    },
+    {
+      "epoch": 2.430674846625767,
+      "grad_norm": 4.089489383771658,
+      "learning_rate": 1.5338842975206612e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -426.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0096,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.4453125,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -11.9375,
+      "step": 1981
+    },
+    {
+      "epoch": 2.4319018404907977,
+      "grad_norm": 1.4287040506933075,
+      "learning_rate": 1.530578512396694e-07,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.93359375,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -12.3125,
+      "step": 1982
+    },
+    {
+      "epoch": 2.4331288343558284,
+      "grad_norm": 1.6905097237452604,
+      "learning_rate": 1.5272727272727273e-07,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -458.0,
+      "logps/rejected": -454.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2021484375,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -12.75,
+      "step": 1983
+    },
+    {
+      "epoch": 2.434355828220859,
+      "grad_norm": 3.0773741827971595,
+      "learning_rate": 1.5239669421487602e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -442.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.06005859375,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.5,
+      "step": 1984
+    },
+    {
+      "epoch": 2.4355828220858897,
+      "grad_norm": 7.08252440776849,
+      "learning_rate": 1.5206611570247934e-07,
+      "logits/chosen": -0.31640625,
+      "logits/rejected": -0.625,
+      "logps/chosen": -454.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0071,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.26171875,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.5625,
+      "step": 1985
+    },
+    {
+      "epoch": 2.4368098159509204,
+      "grad_norm": 1.6155564645929124,
+      "learning_rate": 1.5173553719008263e-07,
+      "logits/chosen": -0.41796875,
+      "logits/rejected": -0.6328125,
+      "logps/chosen": -448.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.466796875,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -12.5,
+      "step": 1986
+    },
+    {
+      "epoch": 2.438036809815951,
+      "grad_norm": 1.883445685993449,
+      "learning_rate": 1.5140495867768595e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5546875,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.375,
+      "step": 1987
+    },
+    {
+      "epoch": 2.4392638036809817,
+      "grad_norm": 1.4580290514576402,
+      "learning_rate": 1.5107438016528924e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.220703125,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.1875,
+      "step": 1988
+    },
+    {
+      "epoch": 2.4404907975460124,
+      "grad_norm": 1.6492481173772273,
+      "learning_rate": 1.5074380165289255e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.296875,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -12.9375,
+      "step": 1989
+    },
+    {
+      "epoch": 2.441717791411043,
+      "grad_norm": 1.642660095607197,
+      "learning_rate": 1.5041322314049587e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.828125,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.75,
+      "step": 1990
+    },
+    {
+      "epoch": 2.4429447852760737,
+      "grad_norm": 1.5316428274651595,
+      "learning_rate": 1.5008264462809916e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.267578125,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.1875,
+      "step": 1991
+    },
+    {
+      "epoch": 2.4441717791411044,
+      "grad_norm": 1.594535895834974,
+      "learning_rate": 1.4975206611570248e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.396484375,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -12.0,
+      "step": 1992
+    },
+    {
+      "epoch": 2.445398773006135,
+      "grad_norm": 4.216608378623086,
+      "learning_rate": 1.4942148760330577e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0071,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2578125,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -11.5625,
+      "step": 1993
+    },
+    {
+      "epoch": 2.4466257668711657,
+      "grad_norm": 4.208056954012315,
+      "learning_rate": 1.490909090909091e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6875,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.375,
+      "step": 1994
+    },
+    {
+      "epoch": 2.4478527607361964,
+      "grad_norm": 2.195936111816115,
+      "learning_rate": 1.4876033057851238e-07,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -458.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -12.0,
+      "step": 1995
+    },
+    {
+      "epoch": 2.449079754601227,
+      "grad_norm": 2.0445108463571158,
+      "learning_rate": 1.484297520661157e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.59375,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -11.875,
+      "step": 1996
+    },
+    {
+      "epoch": 2.4503067484662577,
+      "grad_norm": 4.0861547195504,
+      "learning_rate": 1.48099173553719e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0074,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1748046875,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -11.9375,
+      "step": 1997
+    },
+    {
+      "epoch": 2.4515337423312884,
+      "grad_norm": 1.1875814515273448,
+      "learning_rate": 1.477685950413223e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0025,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1826171875,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -11.8125,
+      "step": 1998
+    },
+    {
+      "epoch": 2.452760736196319,
+      "grad_norm": 1.4176326039352591,
+      "learning_rate": 1.4743801652892563e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0104,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.271484375,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -11.125,
+      "step": 1999
+    },
+    {
+      "epoch": 2.4539877300613497,
+      "grad_norm": 3.4876624531075726,
+      "learning_rate": 1.4710743801652892e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0088,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5234375,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -12.1875,
+      "step": 2000
+    },
+    {
+      "epoch": 2.4552147239263804,
+      "grad_norm": 2.5303296897929894,
+      "learning_rate": 1.4677685950413223e-07,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.03564453125,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -11.9375,
+      "step": 2001
+    },
+    {
+      "epoch": 2.456441717791411,
+      "grad_norm": 1.5501188119885745,
+      "learning_rate": 1.4644628099173553e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.080078125,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.0,
+      "step": 2002
+    },
+    {
+      "epoch": 2.4576687116564417,
+      "grad_norm": 2.8946866155270867,
+      "learning_rate": 1.4611570247933884e-07,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0077,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.63671875,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -11.875,
+      "step": 2003
+    },
+    {
+      "epoch": 2.4588957055214724,
+      "grad_norm": 3.3602826638957772,
+      "learning_rate": 1.4578512396694214e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -438.0,
+      "loss": 0.008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2578125,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -11.75,
+      "step": 2004
+    },
+    {
+      "epoch": 2.460122699386503,
+      "grad_norm": 1.5497608926223454,
+      "learning_rate": 1.4545454545454545e-07,
+      "logits/chosen": -0.404296875,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.232421875,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.125,
+      "step": 2005
+    },
+    {
+      "epoch": 2.4613496932515337,
+      "grad_norm": 2.476304869120195,
+      "learning_rate": 1.4512396694214874e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.298828125,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.0625,
+      "step": 2006
+    },
+    {
+      "epoch": 2.4625766871165644,
+      "grad_norm": 2.8122671505751726,
+      "learning_rate": 1.4479338842975206e-07,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0068,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.451171875,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.75,
+      "step": 2007
+    },
+    {
+      "epoch": 2.463803680981595,
+      "grad_norm": 2.413379933844493,
+      "learning_rate": 1.4446280991735535e-07,
+      "logits/chosen": -0.380859375,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -354.0,
+      "loss": 0.0064,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1796875,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -12.1875,
+      "step": 2008
+    },
+    {
+      "epoch": 2.4650306748466257,
+      "grad_norm": 1.3911644626611408,
+      "learning_rate": 1.4413223140495867e-07,
+      "logits/chosen": -0.4140625,
+      "logits/rejected": -0.64453125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -388.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.51953125,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -13.625,
+      "step": 2009
+    },
+    {
+      "epoch": 2.4662576687116564,
+      "grad_norm": 6.111852212449521,
+      "learning_rate": 1.43801652892562e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0112,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.447265625,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.3125,
+      "step": 2010
+    },
+    {
+      "epoch": 2.467484662576687,
+      "grad_norm": 2.9602726667815493,
+      "learning_rate": 1.4347107438016528e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.7578125,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -11.75,
+      "step": 2011
+    },
+    {
+      "epoch": 2.4687116564417177,
+      "grad_norm": 4.533697164764621,
+      "learning_rate": 1.431404958677686e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0082,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.59765625,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.1875,
+      "step": 2012
+    },
+    {
+      "epoch": 2.4699386503067484,
+      "grad_norm": 2.0151966397209593,
+      "learning_rate": 1.428099173553719e-07,
+      "logits/chosen": -0.451171875,
+      "logits/rejected": -0.6015625,
+      "logps/chosen": -430.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.921875,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -13.9375,
+      "step": 2013
+    },
+    {
+      "epoch": 2.471165644171779,
+      "grad_norm": 2.3119760627516417,
+      "learning_rate": 1.424793388429752e-07,
+      "logits/chosen": -0.357421875,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -444.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.078125,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.75,
+      "step": 2014
+    },
+    {
+      "epoch": 2.4723926380368098,
+      "grad_norm": 1.1854957464045532,
+      "learning_rate": 1.421487603305785e-07,
+      "logits/chosen": -0.3984375,
+      "logits/rejected": -0.6484375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0024,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.08251953125,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.5,
+      "step": 2015
+    },
+    {
+      "epoch": 2.4736196319018404,
+      "grad_norm": 1.4535722798694182,
+      "learning_rate": 1.4181818181818182e-07,
+      "logits/chosen": -0.37890625,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -450.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0086,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.05322265625,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.5625,
+      "step": 2016
+    },
+    {
+      "epoch": 2.474846625766871,
+      "grad_norm": 1.8444079472627,
+      "learning_rate": 1.414876033057851e-07,
+      "logits/chosen": -0.330078125,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.85546875,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.5625,
+      "step": 2017
+    },
+    {
+      "epoch": 2.4760736196319018,
+      "grad_norm": 2.196137610194229,
+      "learning_rate": 1.4115702479338842e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0038,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.703125,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.1875,
+      "step": 2018
+    },
+    {
+      "epoch": 2.4773006134969324,
+      "grad_norm": 2.305337778670854,
+      "learning_rate": 1.4082644628099174e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0057,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8125,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -13.1875,
+      "step": 2019
+    },
+    {
+      "epoch": 2.478527607361963,
+      "grad_norm": 2.673102218140307,
+      "learning_rate": 1.4049586776859503e-07,
+      "logits/chosen": -0.421875,
+      "logits/rejected": -0.62109375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.87890625,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -13.0625,
+      "step": 2020
+    },
+    {
+      "epoch": 2.4797546012269938,
+      "grad_norm": 1.9076369448135233,
+      "learning_rate": 1.4016528925619835e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.408203125,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.8125,
+      "step": 2021
+    },
+    {
+      "epoch": 2.4809815950920244,
+      "grad_norm": 2.721115444947314,
+      "learning_rate": 1.3983471074380164e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0078,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.59375,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -12.1875,
+      "step": 2022
+    },
+    {
+      "epoch": 2.482208588957055,
+      "grad_norm": 1.13647574626614,
+      "learning_rate": 1.3950413223140496e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0029,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.37890625,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.25,
+      "step": 2023
+    },
+    {
+      "epoch": 2.4834355828220858,
+      "grad_norm": 1.367758636969539,
+      "learning_rate": 1.3917355371900825e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -448.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1181640625,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.9375,
+      "step": 2024
+    },
+    {
+      "epoch": 2.4846625766871164,
+      "grad_norm": 2.5296097256989287,
+      "learning_rate": 1.3884297520661157e-07,
+      "logits/chosen": -0.416015625,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0091,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.96484375,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.6875,
+      "step": 2025
+    },
+    {
+      "epoch": 2.485889570552147,
+      "grad_norm": 2.117765864812374,
+      "learning_rate": 1.3851239669421486e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -426.0,
+      "loss": 0.005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.80078125,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.375,
+      "step": 2026
+    },
+    {
+      "epoch": 2.4871165644171778,
+      "grad_norm": 2.1385172238281407,
+      "learning_rate": 1.3818181818181818e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.57421875,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -12.9375,
+      "step": 2027
+    },
+    {
+      "epoch": 2.4883435582822084,
+      "grad_norm": 10.573834083228371,
+      "learning_rate": 1.3785123966942147e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0104,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.337890625,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.5625,
+      "step": 2028
+    },
+    {
+      "epoch": 2.489570552147239,
+      "grad_norm": 1.70762761839659,
+      "learning_rate": 1.3752066115702479e-07,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -426.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0078125,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -12.4375,
+      "step": 2029
+    },
+    {
+      "epoch": 2.4907975460122698,
+      "grad_norm": 2.2150665693262424,
+      "learning_rate": 1.371900826446281e-07,
+      "logits/chosen": -0.376953125,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.271484375,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.0,
+      "step": 2030
+    },
+    {
+      "epoch": 2.4920245398773004,
+      "grad_norm": 2.9338053721800743,
+      "learning_rate": 1.368595041322314e-07,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0077,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.62890625,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -12.125,
+      "step": 2031
+    },
+    {
+      "epoch": 2.493251533742331,
+      "grad_norm": 3.1052081503827975,
+      "learning_rate": 1.365289256198347e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.53125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.75,
+      "step": 2032
+    },
+    {
+      "epoch": 2.4944785276073618,
+      "grad_norm": 2.431190624645661,
+      "learning_rate": 1.36198347107438e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0068,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.01953125,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.0,
+      "step": 2033
+    },
+    {
+      "epoch": 2.4957055214723924,
+      "grad_norm": 1.2787869224653277,
+      "learning_rate": 1.3586776859504132e-07,
+      "logits/chosen": -0.345703125,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -454.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.29296875,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -13.4375,
+      "step": 2034
+    },
+    {
+      "epoch": 2.4969325153374236,
+      "grad_norm": 2.271342891021494,
+      "learning_rate": 1.355371900826446e-07,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.267578125,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.375,
+      "step": 2035
+    },
+    {
+      "epoch": 2.4981595092024538,
+      "grad_norm": 1.2492181772637767,
+      "learning_rate": 1.3520661157024793e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0028,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8515625,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -13.0,
+      "step": 2036
+    },
+    {
+      "epoch": 2.499386503067485,
+      "grad_norm": 2.027702498909458,
+      "learning_rate": 1.3487603305785122e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0108,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.162109375,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.6875,
+      "step": 2037
+    },
+    {
+      "epoch": 2.500613496932515,
+      "grad_norm": 11.996194810594588,
+      "learning_rate": 1.3454545454545454e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0236,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -1.25,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -13.25,
+      "step": 2038
+    },
+    {
+      "epoch": 2.5018404907975462,
+      "grad_norm": 1.4637965501132189,
+      "learning_rate": 1.3421487603305786e-07,
+      "logits/chosen": -0.24609375,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.93359375,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.375,
+      "step": 2039
+    },
+    {
+      "epoch": 2.5030674846625764,
+      "grad_norm": 2.0685232062862378,
+      "learning_rate": 1.3388429752066115e-07,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.71875,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.5,
+      "step": 2040
+    },
+    {
+      "epoch": 2.5042944785276076,
+      "grad_norm": 0.9001043644617583,
+      "learning_rate": 1.3355371900826447e-07,
+      "logits/chosen": -0.419921875,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1572265625,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.3125,
+      "step": 2041
+    },
+    {
+      "epoch": 2.505521472392638,
+      "grad_norm": 2.7711300372842063,
+      "learning_rate": 1.3322314049586776e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0056,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.263671875,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -12.75,
+      "step": 2042
+    },
+    {
+      "epoch": 2.506748466257669,
+      "grad_norm": 2.303112129378537,
+      "learning_rate": 1.3289256198347108e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2890625,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.6875,
+      "step": 2043
+    },
+    {
+      "epoch": 2.507975460122699,
+      "grad_norm": 6.047374137694716,
+      "learning_rate": 1.3256198347107437e-07,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -318.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0149,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.86328125,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.5625,
+      "step": 2044
+    },
+    {
+      "epoch": 2.5092024539877302,
+      "grad_norm": 1.6838688284037515,
+      "learning_rate": 1.3223140495867768e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.283203125,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.0,
+      "step": 2045
+    },
+    {
+      "epoch": 2.510429447852761,
+      "grad_norm": 1.598144447488392,
+      "learning_rate": 1.3190082644628098e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.68359375,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -12.1875,
+      "step": 2046
+    },
+    {
+      "epoch": 2.5116564417177916,
+      "grad_norm": 3.599895856570797,
+      "learning_rate": 1.315702479338843e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0159,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.51171875,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.875,
+      "step": 2047
+    },
+    {
+      "epoch": 2.5128834355828222,
+      "grad_norm": 4.3841941777268145,
+      "learning_rate": 1.3123966942148758e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0104,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.7109375,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.5625,
+      "step": 2048
+    },
+    {
+      "epoch": 2.514110429447853,
+      "grad_norm": 1.9631397371011616,
+      "learning_rate": 1.309090909090909e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -456.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0085,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.37109375,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -12.3125,
+      "step": 2049
+    },
+    {
+      "epoch": 2.5153374233128836,
+      "grad_norm": 2.102157379411208,
+      "learning_rate": 1.3057851239669422e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -458.0,
+      "logps/rejected": -440.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.322265625,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.5,
+      "step": 2050
+    },
+    {
+      "epoch": 2.5165644171779142,
+      "grad_norm": 1.8968324060941286,
+      "learning_rate": 1.302479338842975e-07,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.5,
+      "logps/chosen": -422.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1396484375,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -11.5625,
+      "step": 2051
+    },
+    {
+      "epoch": 2.517791411042945,
+      "grad_norm": 4.019978570137844,
+      "learning_rate": 1.2991735537190083e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0156,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.5625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -12.125,
+      "step": 2052
+    },
+    {
+      "epoch": 2.5190184049079756,
+      "grad_norm": 2.1569151771318955,
+      "learning_rate": 1.2958677685950412e-07,
+      "logits/chosen": -0.390625,
+      "logits/rejected": -0.5546875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.640625,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -12.1875,
+      "step": 2053
+    },
+    {
+      "epoch": 2.5202453987730062,
+      "grad_norm": 1.3178633904085888,
+      "learning_rate": 1.2925619834710744e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -456.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0179443359375,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.875,
+      "step": 2054
+    },
+    {
+      "epoch": 2.521472392638037,
+      "grad_norm": 1.1562371642342075,
+      "learning_rate": 1.2892561983471073e-07,
+      "logits/chosen": -0.10107421875,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6796875,
+      "rewards/margins": 11.125,
+      "rewards/rejected": -11.8125,
+      "step": 2055
+    },
+    {
+      "epoch": 2.5226993865030676,
+      "grad_norm": 4.660491971791105,
+      "learning_rate": 1.2859504132231405e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -452.0,
+      "logps/rejected": -440.0,
+      "loss": 0.007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.57421875,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.125,
+      "step": 2056
+    },
+    {
+      "epoch": 2.5239263803680982,
+      "grad_norm": 1.9606321508504119,
+      "learning_rate": 1.2826446280991734e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -420.0,
+      "logps/rejected": -440.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.019775390625,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.75,
+      "step": 2057
+    },
+    {
+      "epoch": 2.525153374233129,
+      "grad_norm": 1.5432004743805674,
+      "learning_rate": 1.2793388429752066e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -438.0,
+      "logps/rejected": -454.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2138671875,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -12.8125,
+      "step": 2058
+    },
+    {
+      "epoch": 2.5263803680981596,
+      "grad_norm": 2.8205561578125935,
+      "learning_rate": 1.2760330578512397e-07,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.11279296875,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.6875,
+      "step": 2059
+    },
+    {
+      "epoch": 2.5276073619631902,
+      "grad_norm": 1.8607015440154648,
+      "learning_rate": 1.2727272727272726e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.33984375,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.4375,
+      "step": 2060
+    },
+    {
+      "epoch": 2.528834355828221,
+      "grad_norm": 2.2532130000905055,
+      "learning_rate": 1.2694214876033058e-07,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.15625,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -13.0625,
+      "step": 2061
+    },
+    {
+      "epoch": 2.5300613496932516,
+      "grad_norm": 1.0029768938599994,
+      "learning_rate": 1.2661157024793387e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0025,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.01300048828125,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.875,
+      "step": 2062
+    },
+    {
+      "epoch": 2.5312883435582823,
+      "grad_norm": 2.681130201983871,
+      "learning_rate": 1.262809917355372e-07,
+      "logits/chosen": -0.447265625,
+      "logits/rejected": -0.6015625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0038,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5234375,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.1875,
+      "step": 2063
+    },
+    {
+      "epoch": 2.532515337423313,
+      "grad_norm": 1.131646759706482,
+      "learning_rate": 1.2595041322314048e-07,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.63671875,
+      "logps/chosen": -440.0,
+      "logps/rejected": -432.0,
+      "loss": 0.0029,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.001800537109375,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.75,
+      "step": 2064
+    },
+    {
+      "epoch": 2.5337423312883436,
+      "grad_norm": 2.300337569573409,
+      "learning_rate": 1.256198347107438e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0075,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.640625,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -12.125,
+      "step": 2065
+    },
+    {
+      "epoch": 2.5349693251533743,
+      "grad_norm": 4.268030562873055,
+      "learning_rate": 1.252892561983471e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0109,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.73828125,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -12.1875,
+      "step": 2066
+    },
+    {
+      "epoch": 2.536196319018405,
+      "grad_norm": 1.522099180147848,
+      "learning_rate": 1.249586776859504e-07,
+      "logits/chosen": -0.365234375,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -408.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8359375,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.9375,
+      "step": 2067
+    },
+    {
+      "epoch": 2.5374233128834356,
+      "grad_norm": 1.5639024130282178,
+      "learning_rate": 1.246280991735537e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.57421875,
+      "logps/chosen": -462.0,
+      "logps/rejected": -426.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.4609375,
+      "rewards/margins": 14.0,
+      "rewards/rejected": -13.5,
+      "step": 2068
+    },
+    {
+      "epoch": 2.5386503067484663,
+      "grad_norm": 2.055776856761035,
+      "learning_rate": 1.2429752066115702e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.173828125,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -13.125,
+      "step": 2069
+    },
+    {
+      "epoch": 2.539877300613497,
+      "grad_norm": 1.2615291646278584,
+      "learning_rate": 1.2396694214876034e-07,
+      "logits/chosen": -0.380859375,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.578125,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.75,
+      "step": 2070
+    },
+    {
+      "epoch": 2.5411042944785276,
+      "grad_norm": 1.391494474203479,
+      "learning_rate": 1.2363636363636363e-07,
+      "logits/chosen": -0.36328125,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -436.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.283203125,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -12.75,
+      "step": 2071
+    },
+    {
+      "epoch": 2.5423312883435583,
+      "grad_norm": 1.2419768744499535,
+      "learning_rate": 1.2330578512396694e-07,
+      "logits/chosen": -0.353515625,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0029,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.38671875,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.375,
+      "step": 2072
+    },
+    {
+      "epoch": 2.543558282208589,
+      "grad_norm": 1.689735711552667,
+      "learning_rate": 1.2297520661157024e-07,
+      "logits/chosen": -0.412109375,
+      "logits/rejected": -0.6171875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.15234375,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -13.0625,
+      "step": 2073
+    },
+    {
+      "epoch": 2.5447852760736196,
+      "grad_norm": 1.1398112927489716,
+      "learning_rate": 1.2264462809917355e-07,
+      "logits/chosen": -0.369140625,
+      "logits/rejected": -0.6796875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0079,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.453125,
+      "rewards/margins": 13.4375,
+      "rewards/rejected": -13.0,
+      "step": 2074
+    },
+    {
+      "epoch": 2.5460122699386503,
+      "grad_norm": 4.344352004176359,
+      "learning_rate": 1.2231404958677684e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0143,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.29296875,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.3125,
+      "step": 2075
+    },
+    {
+      "epoch": 2.547239263803681,
+      "grad_norm": 2.93772666145552,
+      "learning_rate": 1.2198347107438016e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.466796875,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.6875,
+      "step": 2076
+    },
+    {
+      "epoch": 2.5484662576687116,
+      "grad_norm": 1.263814528413439,
+      "learning_rate": 1.2165289256198345e-07,
+      "logits/chosen": -0.421875,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -460.0,
+      "logps/rejected": -450.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2177734375,
+      "rewards/margins": 13.5625,
+      "rewards/rejected": -13.8125,
+      "step": 2077
+    },
+    {
+      "epoch": 2.5496932515337423,
+      "grad_norm": 1.4830372189681982,
+      "learning_rate": 1.2132231404958677e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -394.0,
+      "loss": 0.009,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.0703125,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -12.3125,
+      "step": 2078
+    },
+    {
+      "epoch": 2.550920245398773,
+      "grad_norm": 1.8890603608625867,
+      "learning_rate": 1.209917355371901e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.283203125,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.5625,
+      "step": 2079
+    },
+    {
+      "epoch": 2.5521472392638036,
+      "grad_norm": 4.820957161396197,
+      "learning_rate": 1.2066115702479338e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0071,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.453125,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.0,
+      "step": 2080
+    },
+    {
+      "epoch": 2.5533742331288343,
+      "grad_norm": 1.125274987639882,
+      "learning_rate": 1.203305785123967e-07,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.53515625,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -13.5625,
+      "step": 2081
+    },
+    {
+      "epoch": 2.554601226993865,
+      "grad_norm": 2.9030438161501833,
+      "learning_rate": 1.2e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.373046875,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -12.625,
+      "step": 2082
+    },
+    {
+      "epoch": 2.5558282208588956,
+      "grad_norm": 2.5555453195092825,
+      "learning_rate": 1.196694214876033e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0127,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.140625,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -12.25,
+      "step": 2083
+    },
+    {
+      "epoch": 2.5570552147239263,
+      "grad_norm": 1.8655554317698846,
+      "learning_rate": 1.193388429752066e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0098,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.203125,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -13.0,
+      "step": 2084
+    },
+    {
+      "epoch": 2.558282208588957,
+      "grad_norm": 2.2393032665920596,
+      "learning_rate": 1.1900826446280992e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.76953125,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -12.0,
+      "step": 2085
+    },
+    {
+      "epoch": 2.5595092024539876,
+      "grad_norm": 1.8284503634433844,
+      "learning_rate": 1.1867768595041321e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.11474609375,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.6875,
+      "step": 2086
+    },
+    {
+      "epoch": 2.5607361963190183,
+      "grad_norm": 1.098939263693094,
+      "learning_rate": 1.1834710743801652e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -13.375,
+      "step": 2087
+    },
+    {
+      "epoch": 2.561963190184049,
+      "grad_norm": 0.7894340963601841,
+      "learning_rate": 1.1801652892561984e-07,
+      "logits/chosen": -0.400390625,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0021,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.470703125,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -13.0,
+      "step": 2088
+    },
+    {
+      "epoch": 2.5631901840490796,
+      "grad_norm": 1.7102188828800173,
+      "learning_rate": 1.1768595041322313e-07,
+      "logits/chosen": -0.3828125,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0056,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.75390625,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -13.6875,
+      "step": 2089
+    },
+    {
+      "epoch": 2.5644171779141103,
+      "grad_norm": 2.544731803356017,
+      "learning_rate": 1.1735537190082645e-07,
+      "logits/chosen": -0.439453125,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0062,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.53125,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -13.875,
+      "step": 2090
+    },
+    {
+      "epoch": 2.565644171779141,
+      "grad_norm": 1.382262030125881,
+      "learning_rate": 1.1702479338842974e-07,
+      "logits/chosen": -0.431640625,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0038,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.70703125,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.625,
+      "step": 2091
+    },
+    {
+      "epoch": 2.5668711656441716,
+      "grad_norm": 1.4994101123615182,
+      "learning_rate": 1.1669421487603306e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -456.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.287109375,
+      "rewards/margins": 13.875,
+      "rewards/rejected": -13.625,
+      "step": 2092
+    },
+    {
+      "epoch": 2.5680981595092023,
+      "grad_norm": 0.9568856582345134,
+      "learning_rate": 1.1636363636363635e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0023,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1943359375,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -12.8125,
+      "step": 2093
+    },
+    {
+      "epoch": 2.5693251533742334,
+      "grad_norm": 3.6453808217581685,
+      "learning_rate": 1.1603305785123967e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0062,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6484375,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.6875,
+      "step": 2094
+    },
+    {
+      "epoch": 2.5705521472392636,
+      "grad_norm": 4.774821867202303,
+      "learning_rate": 1.1570247933884296e-07,
+      "logits/chosen": -0.369140625,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0098,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.125,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.1875,
+      "step": 2095
+    },
+    {
+      "epoch": 2.5717791411042947,
+      "grad_norm": 1.226088325462221,
+      "learning_rate": 1.1537190082644628e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.609375,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.875,
+      "step": 2096
+    },
+    {
+      "epoch": 2.573006134969325,
+      "grad_norm": 1.7596810367892974,
+      "learning_rate": 1.1504132231404957e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -432.0,
+      "loss": 0.0068,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5078125,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.5,
+      "step": 2097
+    },
+    {
+      "epoch": 2.574233128834356,
+      "grad_norm": 3.0063572525817515,
+      "learning_rate": 1.1471074380165289e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.734375,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -12.3125,
+      "step": 2098
+    },
+    {
+      "epoch": 2.5754601226993863,
+      "grad_norm": 6.89183831600462,
+      "learning_rate": 1.143801652892562e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0066,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.333984375,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -12.75,
+      "step": 2099
+    },
+    {
+      "epoch": 2.5766871165644174,
+      "grad_norm": 1.108029401793295,
+      "learning_rate": 1.140495867768595e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0024,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.09765625,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -12.9375,
+      "step": 2100
+    },
+    {
+      "epoch": 2.5779141104294476,
+      "grad_norm": 0.6966779815695421,
+      "learning_rate": 1.1371900826446281e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0017,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.59765625,
+      "rewards/margins": 13.1875,
+      "rewards/rejected": -12.5625,
+      "step": 2101
+    },
+    {
+      "epoch": 2.5791411042944787,
+      "grad_norm": 1.6191064315182209,
+      "learning_rate": 1.133884297520661e-07,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -360.0,
+      "logps/rejected": -384.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.21484375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.375,
+      "step": 2102
+    },
+    {
+      "epoch": 2.580368098159509,
+      "grad_norm": 1.782654269639726,
+      "learning_rate": 1.1305785123966942e-07,
+      "logits/chosen": -0.2001953125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.039306640625,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.25,
+      "step": 2103
+    },
+    {
+      "epoch": 2.58159509202454,
+      "grad_norm": 1.676280045154247,
+      "learning_rate": 1.1272727272727271e-07,
+      "logits/chosen": -0.0625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.828125,
+      "rewards/margins": 10.5,
+      "rewards/rejected": -11.3125,
+      "step": 2104
+    },
+    {
+      "epoch": 2.5828220858895703,
+      "grad_norm": 1.1929843951960957,
+      "learning_rate": 1.1239669421487603e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.08349609375,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.375,
+      "step": 2105
+    },
+    {
+      "epoch": 2.5840490797546014,
+      "grad_norm": 2.079807612894869,
+      "learning_rate": 1.1206611570247932e-07,
+      "logits/chosen": -0.404296875,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.65234375,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.625,
+      "step": 2106
+    },
+    {
+      "epoch": 2.5852760736196316,
+      "grad_norm": 2.4913147679513123,
+      "learning_rate": 1.1173553719008264e-07,
+      "logits/chosen": -0.41015625,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1640625,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -13.0625,
+      "step": 2107
+    },
+    {
+      "epoch": 2.5865030674846627,
+      "grad_norm": 1.138008200011956,
+      "learning_rate": 1.1140495867768596e-07,
+      "logits/chosen": -0.37109375,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -410.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.318359375,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.1875,
+      "step": 2108
+    },
+    {
+      "epoch": 2.5877300613496934,
+      "grad_norm": 2.7457707744360467,
+      "learning_rate": 1.1107438016528925e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -400.0,
+      "loss": 0.005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.65234375,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -11.75,
+      "step": 2109
+    },
+    {
+      "epoch": 2.588957055214724,
+      "grad_norm": 1.9931457416455565,
+      "learning_rate": 1.1074380165289257e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1669921875,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.25,
+      "step": 2110
+    },
+    {
+      "epoch": 2.5901840490797547,
+      "grad_norm": 2.3431096055817933,
+      "learning_rate": 1.1041322314049586e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -348.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.76953125,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -11.5625,
+      "step": 2111
+    },
+    {
+      "epoch": 2.5914110429447854,
+      "grad_norm": 1.0575368546986408,
+      "learning_rate": 1.1008264462809918e-07,
+      "logits/chosen": -0.1162109375,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.296875,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.5,
+      "step": 2112
+    },
+    {
+      "epoch": 2.592638036809816,
+      "grad_norm": 1.6528555772459397,
+      "learning_rate": 1.0975206611570247e-07,
+      "logits/chosen": -0.248046875,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.98828125,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -12.625,
+      "step": 2113
+    },
+    {
+      "epoch": 2.5938650306748468,
+      "grad_norm": 1.936922176920219,
+      "learning_rate": 1.0942148760330579e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0830078125,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -11.8125,
+      "step": 2114
+    },
+    {
+      "epoch": 2.5950920245398774,
+      "grad_norm": 2.4746167878752257,
+      "learning_rate": 1.0909090909090908e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0112,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.447265625,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -11.8125,
+      "step": 2115
+    },
+    {
+      "epoch": 2.596319018404908,
+      "grad_norm": 1.6835171337696344,
+      "learning_rate": 1.087603305785124e-07,
+      "logits/chosen": -0.126953125,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.365234375,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -10.875,
+      "step": 2116
+    },
+    {
+      "epoch": 2.5975460122699388,
+      "grad_norm": 1.2780023236362608,
+      "learning_rate": 1.0842975206611569e-07,
+      "logits/chosen": -0.345703125,
+      "logits/rejected": -0.57421875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.125,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.0,
+      "step": 2117
+    },
+    {
+      "epoch": 2.5987730061349694,
+      "grad_norm": 1.89116145530193,
+      "learning_rate": 1.08099173553719e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -376.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.546875,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.1875,
+      "step": 2118
+    },
+    {
+      "epoch": 2.6,
+      "grad_norm": 1.447506037595178,
+      "learning_rate": 1.0776859504132232e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0089,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.498046875,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.3125,
+      "step": 2119
+    },
+    {
+      "epoch": 2.6012269938650308,
+      "grad_norm": 1.8215159424918526,
+      "learning_rate": 1.0743801652892561e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.14453125,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.0,
+      "step": 2120
+    },
+    {
+      "epoch": 2.6024539877300614,
+      "grad_norm": 1.9064672806580176,
+      "learning_rate": 1.0710743801652893e-07,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.037353515625,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -12.6875,
+      "step": 2121
+    },
+    {
+      "epoch": 2.603680981595092,
+      "grad_norm": 1.4958145389556794,
+      "learning_rate": 1.0677685950413222e-07,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0087,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.318359375,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.0,
+      "step": 2122
+    },
+    {
+      "epoch": 2.6049079754601228,
+      "grad_norm": 1.8550826589528993,
+      "learning_rate": 1.0644628099173554e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0091,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.12255859375,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -11.75,
+      "step": 2123
+    },
+    {
+      "epoch": 2.6061349693251534,
+      "grad_norm": 11.74094882017673,
+      "learning_rate": 1.0611570247933883e-07,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.55078125,
+      "logps/chosen": -420.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0096,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.78515625,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -12.5,
+      "step": 2124
+    },
+    {
+      "epoch": 2.607361963190184,
+      "grad_norm": 0.9347077058844525,
+      "learning_rate": 1.0578512396694215e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -462.0,
+      "loss": 0.0075,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.30078125,
+      "rewards/margins": 13.1875,
+      "rewards/rejected": -12.875,
+      "step": 2125
+    },
+    {
+      "epoch": 2.6085889570552148,
+      "grad_norm": 2.1474283243273233,
+      "learning_rate": 1.0545454545454544e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0673828125,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -11.9375,
+      "step": 2126
+    },
+    {
+      "epoch": 2.6098159509202454,
+      "grad_norm": 2.3889484461255424,
+      "learning_rate": 1.0512396694214876e-07,
+      "logits/chosen": -0.40234375,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.71875,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -12.375,
+      "step": 2127
+    },
+    {
+      "epoch": 2.611042944785276,
+      "grad_norm": 2.563824467263969,
+      "learning_rate": 1.0479338842975207e-07,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -464.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.21875,
+      "rewards/margins": 13.4375,
+      "rewards/rejected": -13.625,
+      "step": 2128
+    },
+    {
+      "epoch": 2.6122699386503068,
+      "grad_norm": 2.5154361575291833,
+      "learning_rate": 1.0446280991735537e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.5546875,
+      "logps/chosen": -438.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.0050048828125,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -12.9375,
+      "step": 2129
+    },
+    {
+      "epoch": 2.6134969325153374,
+      "grad_norm": 5.452655663166473,
+      "learning_rate": 1.0413223140495868e-07,
+      "logits/chosen": -0.3828125,
+      "logits/rejected": -0.62890625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0084,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.546875,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -13.0625,
+      "step": 2130
+    },
+    {
+      "epoch": 2.614723926380368,
+      "grad_norm": 1.9383177281147583,
+      "learning_rate": 1.0380165289256197e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.3828125,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.8125,
+      "step": 2131
+    },
+    {
+      "epoch": 2.6159509202453988,
+      "grad_norm": 1.6774518072508986,
+      "learning_rate": 1.0347107438016529e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -446.0,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1298828125,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.3125,
+      "step": 2132
+    },
+    {
+      "epoch": 2.6171779141104294,
+      "grad_norm": 2.9100827479820737,
+      "learning_rate": 1.0314049586776858e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.435546875,
+      "rewards/margins": 13.5,
+      "rewards/rejected": -13.0625,
+      "step": 2133
+    },
+    {
+      "epoch": 2.61840490797546,
+      "grad_norm": 1.4962937176620514,
+      "learning_rate": 1.028099173553719e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.5546875,
+      "logps/chosen": -438.0,
+      "logps/rejected": -452.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.00201416015625,
+      "rewards/margins": 13.1875,
+      "rewards/rejected": -13.1875,
+      "step": 2134
+    },
+    {
+      "epoch": 2.6196319018404908,
+      "grad_norm": 2.1367739628011204,
+      "learning_rate": 1.0247933884297519e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0703125,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.8125,
+      "step": 2135
+    },
+    {
+      "epoch": 2.6208588957055214,
+      "grad_norm": 1.917233968240826,
+      "learning_rate": 1.0214876033057851e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.640625,
+      "logps/chosen": -476.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.388671875,
+      "rewards/margins": 13.625,
+      "rewards/rejected": -13.25,
+      "step": 2136
+    },
+    {
+      "epoch": 2.622085889570552,
+      "grad_norm": 2.0301943828988107,
+      "learning_rate": 1.018181818181818e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.27734375,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -12.9375,
+      "step": 2137
+    },
+    {
+      "epoch": 2.6233128834355828,
+      "grad_norm": 2.652938627360609,
+      "learning_rate": 1.0148760330578512e-07,
+      "logits/chosen": -0.41796875,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0062,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.43359375,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.0625,
+      "step": 2138
+    },
+    {
+      "epoch": 2.6245398773006134,
+      "grad_norm": 1.1928229342448688,
+      "learning_rate": 1.0115702479338844e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0025,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.130859375,
+      "rewards/margins": 13.625,
+      "rewards/rejected": -13.75,
+      "step": 2139
+    },
+    {
+      "epoch": 2.625766871165644,
+      "grad_norm": 2.1867122481092656,
+      "learning_rate": 1.0082644628099173e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.03125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.25,
+      "step": 2140
+    },
+    {
+      "epoch": 2.626993865030675,
+      "grad_norm": 1.8384056535563202,
+      "learning_rate": 1.0049586776859505e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.76953125,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.875,
+      "step": 2141
+    },
+    {
+      "epoch": 2.6282208588957054,
+      "grad_norm": 9.861316359127873,
+      "learning_rate": 1.0016528925619834e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0358,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.50390625,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.25,
+      "step": 2142
+    },
+    {
+      "epoch": 2.629447852760736,
+      "grad_norm": 2.7522519700142385,
+      "learning_rate": 9.983471074380165e-08,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.296875,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.75,
+      "step": 2143
+    },
+    {
+      "epoch": 2.630674846625767,
+      "grad_norm": 1.4063257243422123,
+      "learning_rate": 9.950413223140496e-08,
+      "logits/chosen": -0.35546875,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.75390625,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.8125,
+      "step": 2144
+    },
+    {
+      "epoch": 2.6319018404907975,
+      "grad_norm": 4.519758173463782,
+      "learning_rate": 9.917355371900826e-08,
+      "logits/chosen": -0.107421875,
+      "logits/rejected": -0.25,
+      "logps/chosen": -380.0,
+      "logps/rejected": -382.0,
+      "loss": 0.007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1572265625,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.6875,
+      "step": 2145
+    },
+    {
+      "epoch": 2.633128834355828,
+      "grad_norm": 4.731051232562636,
+      "learning_rate": 9.884297520661157e-08,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0116,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6640625,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -13.0,
+      "step": 2146
+    },
+    {
+      "epoch": 2.634355828220859,
+      "grad_norm": 1.1826913207084568,
+      "learning_rate": 9.851239669421487e-08,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0028,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.76171875,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.8125,
+      "step": 2147
+    },
+    {
+      "epoch": 2.6355828220858895,
+      "grad_norm": 3.2085621854602606,
+      "learning_rate": 9.818181818181818e-08,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0097,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.90625,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -12.0625,
+      "step": 2148
+    },
+    {
+      "epoch": 2.63680981595092,
+      "grad_norm": 1.556918625765956,
+      "learning_rate": 9.785123966942148e-08,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.326171875,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.5,
+      "step": 2149
+    },
+    {
+      "epoch": 2.638036809815951,
+      "grad_norm": 4.0893756328358535,
+      "learning_rate": 9.752066115702479e-08,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0071,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.234375,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -11.5625,
+      "step": 2150
+    },
+    {
+      "epoch": 2.6392638036809815,
+      "grad_norm": 1.8157750269153823,
+      "learning_rate": 9.719008264462809e-08,
+      "logits/chosen": -0.341796875,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.06494140625,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.9375,
+      "step": 2151
+    },
+    {
+      "epoch": 2.640490797546012,
+      "grad_norm": 2.454501575656243,
+      "learning_rate": 9.68595041322314e-08,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -436.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1630859375,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -11.875,
+      "step": 2152
+    },
+    {
+      "epoch": 2.641717791411043,
+      "grad_norm": 1.134546092753758,
+      "learning_rate": 9.652892561983471e-08,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0029,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.546875,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.5,
+      "step": 2153
+    },
+    {
+      "epoch": 2.6429447852760735,
+      "grad_norm": 2.3125564679019197,
+      "learning_rate": 9.619834710743802e-08,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -434.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.251953125,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -13.25,
+      "step": 2154
+    },
+    {
+      "epoch": 2.644171779141104,
+      "grad_norm": 1.7144774872375566,
+      "learning_rate": 9.586776859504132e-08,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.140625,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -11.75,
+      "step": 2155
+    },
+    {
+      "epoch": 2.645398773006135,
+      "grad_norm": 1.171220844618522,
+      "learning_rate": 9.553719008264463e-08,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.028564453125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.25,
+      "step": 2156
+    },
+    {
+      "epoch": 2.646625766871166,
+      "grad_norm": 1.1834493345923733,
+      "learning_rate": 9.520661157024793e-08,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -12.625,
+      "step": 2157
+    },
+    {
+      "epoch": 2.647852760736196,
+      "grad_norm": 1.6589085877157108,
+      "learning_rate": 9.487603305785123e-08,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.01513671875,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.0,
+      "step": 2158
+    },
+    {
+      "epoch": 2.6490797546012272,
+      "grad_norm": 1.906883682921282,
+      "learning_rate": 9.454545454545454e-08,
+      "logits/chosen": -0.46484375,
+      "logits/rejected": -0.6171875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0234375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -13.1875,
+      "step": 2159
+    },
+    {
+      "epoch": 2.6503067484662575,
+      "grad_norm": 2.437063079057163,
+      "learning_rate": 9.421487603305784e-08,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0302734375,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -11.8125,
+      "step": 2160
+    },
+    {
+      "epoch": 2.6515337423312886,
+      "grad_norm": 1.0032225485849338,
+      "learning_rate": 9.388429752066115e-08,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0075,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.443359375,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.9375,
+      "step": 2161
+    },
+    {
+      "epoch": 2.652760736196319,
+      "grad_norm": 1.6521900286199462,
+      "learning_rate": 9.355371900826445e-08,
+      "logits/chosen": -0.38671875,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -412.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1484375,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.875,
+      "step": 2162
+    },
+    {
+      "epoch": 2.65398773006135,
+      "grad_norm": 2.0828673125641455,
+      "learning_rate": 9.322314049586777e-08,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5546875,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.1875,
+      "step": 2163
+    },
+    {
+      "epoch": 2.65521472392638,
+      "grad_norm": 0.8313400123528686,
+      "learning_rate": 9.289256198347107e-08,
+      "logits/chosen": -0.38671875,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0024,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.06396484375,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.25,
+      "step": 2164
+    },
+    {
+      "epoch": 2.6564417177914113,
+      "grad_norm": 4.044256521909741,
+      "learning_rate": 9.256198347107438e-08,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0072,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.7890625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -12.3125,
+      "step": 2165
+    },
+    {
+      "epoch": 2.6576687116564415,
+      "grad_norm": 1.440318347761183,
+      "learning_rate": 9.223140495867768e-08,
+      "logits/chosen": -0.35546875,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0093,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.423828125,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.9375,
+      "step": 2166
+    },
+    {
+      "epoch": 2.6588957055214726,
+      "grad_norm": 5.057842339148623,
+      "learning_rate": 9.190082644628099e-08,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0064,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.60546875,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -11.875,
+      "step": 2167
+    },
+    {
+      "epoch": 2.660122699386503,
+      "grad_norm": 1.6917084572751824,
+      "learning_rate": 9.157024793388429e-08,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.60546875,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.375,
+      "step": 2168
+    },
+    {
+      "epoch": 2.661349693251534,
+      "grad_norm": 1.0669987026936152,
+      "learning_rate": 9.12396694214876e-08,
+      "logits/chosen": -0.3671875,
+      "logits/rejected": -0.60546875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -442.0,
+      "loss": 0.0077,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.009033203125,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -13.25,
+      "step": 2169
+    },
+    {
+      "epoch": 2.662576687116564,
+      "grad_norm": 3.1406377749611973,
+      "learning_rate": 9.09090909090909e-08,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0119,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.20703125,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.3125,
+      "step": 2170
+    },
+    {
+      "epoch": 2.6638036809815953,
+      "grad_norm": 2.03346395519121,
+      "learning_rate": 9.05785123966942e-08,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.423828125,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -11.875,
+      "step": 2171
+    },
+    {
+      "epoch": 2.665030674846626,
+      "grad_norm": 1.3149501055085726,
+      "learning_rate": 9.024793388429751e-08,
+      "logits/chosen": -0.31640625,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -328.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0089,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.27734375,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.625,
+      "step": 2172
+    },
+    {
+      "epoch": 2.6662576687116566,
+      "grad_norm": 1.4006174330963654,
+      "learning_rate": 8.991735537190083e-08,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0025,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1376953125,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -12.875,
+      "step": 2173
+    },
+    {
+      "epoch": 2.6674846625766873,
+      "grad_norm": 1.6024081235419565,
+      "learning_rate": 8.958677685950413e-08,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0058,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.87109375,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -12.5,
+      "step": 2174
+    },
+    {
+      "epoch": 2.668711656441718,
+      "grad_norm": 1.3110822237807749,
+      "learning_rate": 8.925619834710744e-08,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5390625,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.625,
+      "step": 2175
+    },
+    {
+      "epoch": 2.6699386503067486,
+      "grad_norm": 1.4304107412787301,
+      "learning_rate": 8.892561983471074e-08,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0092,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.59375,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.75,
+      "step": 2176
+    },
+    {
+      "epoch": 2.6711656441717793,
+      "grad_norm": 0.9355662392784919,
+      "learning_rate": 8.859504132231405e-08,
+      "logits/chosen": -0.380859375,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0024,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.31640625,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.0625,
+      "step": 2177
+    },
+    {
+      "epoch": 2.67239263803681,
+      "grad_norm": 2.5535092776165054,
+      "learning_rate": 8.826446280991735e-08,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0071,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6484375,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -11.5625,
+      "step": 2178
+    },
+    {
+      "epoch": 2.6736196319018406,
+      "grad_norm": 2.8663731534487087,
+      "learning_rate": 8.793388429752065e-08,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.75,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.875,
+      "step": 2179
+    },
+    {
+      "epoch": 2.6748466257668713,
+      "grad_norm": 3.6096817743197773,
+      "learning_rate": 8.760330578512396e-08,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0065,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.96484375,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -12.375,
+      "step": 2180
+    },
+    {
+      "epoch": 2.676073619631902,
+      "grad_norm": 2.3748059732465947,
+      "learning_rate": 8.727272727272726e-08,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0057,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.2421875,
+      "rewards/margins": 10.8125,
+      "rewards/rejected": -12.0625,
+      "step": 2181
+    },
+    {
+      "epoch": 2.6773006134969326,
+      "grad_norm": 6.503981039277375,
+      "learning_rate": 8.694214876033057e-08,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0101,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.609375,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.5625,
+      "step": 2182
+    },
+    {
+      "epoch": 2.6785276073619633,
+      "grad_norm": 8.694130676228118,
+      "learning_rate": 8.661157024793389e-08,
+      "logits/chosen": -0.40234375,
+      "logits/rejected": -0.59765625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0117,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.41796875,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -13.4375,
+      "step": 2183
+    },
+    {
+      "epoch": 2.679754601226994,
+      "grad_norm": 1.1701386733892551,
+      "learning_rate": 8.628099173553719e-08,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.82421875,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.9375,
+      "step": 2184
+    },
+    {
+      "epoch": 2.6809815950920246,
+      "grad_norm": 1.6150696271644773,
+      "learning_rate": 8.59504132231405e-08,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.54296875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.94921875,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.4375,
+      "step": 2185
+    },
+    {
+      "epoch": 2.6822085889570553,
+      "grad_norm": 8.076775949723086,
+      "learning_rate": 8.56198347107438e-08,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.2421875,
+      "logps/chosen": -340.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0104,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1953125,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.9375,
+      "step": 2186
+    },
+    {
+      "epoch": 2.683435582822086,
+      "grad_norm": 1.7762225168066181,
+      "learning_rate": 8.52892561983471e-08,
+      "logits/chosen": -0.39453125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.66796875,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -13.625,
+      "step": 2187
+    },
+    {
+      "epoch": 2.6846625766871166,
+      "grad_norm": 3.3045504733182636,
+      "learning_rate": 8.495867768595041e-08,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.490234375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0057,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.703125,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.25,
+      "step": 2188
+    },
+    {
+      "epoch": 2.6858895705521473,
+      "grad_norm": 1.2149168002154807,
+      "learning_rate": 8.462809917355371e-08,
+      "logits/chosen": -0.384765625,
+      "logits/rejected": -0.58203125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.470703125,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -12.875,
+      "step": 2189
+    },
+    {
+      "epoch": 2.687116564417178,
+      "grad_norm": 2.3918842425048807,
+      "learning_rate": 8.429752066115702e-08,
+      "logits/chosen": -0.39453125,
+      "logits/rejected": -0.59765625,
+      "logps/chosen": -434.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0056,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.125,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.9375,
+      "step": 2190
+    },
+    {
+      "epoch": 2.6883435582822086,
+      "grad_norm": 0.8531096269263563,
+      "learning_rate": 8.396694214876032e-08,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0072,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.197265625,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.0,
+      "step": 2191
+    },
+    {
+      "epoch": 2.6895705521472393,
+      "grad_norm": 1.701212349982693,
+      "learning_rate": 8.363636363636363e-08,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -424.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.46875,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.0625,
+      "step": 2192
+    },
+    {
+      "epoch": 2.69079754601227,
+      "grad_norm": 0.8662731400976823,
+      "learning_rate": 8.330578512396694e-08,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0019,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.79296875,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -13.125,
+      "step": 2193
+    },
+    {
+      "epoch": 2.6920245398773006,
+      "grad_norm": 2.662118587056822,
+      "learning_rate": 8.297520661157025e-08,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.60546875,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.875,
+      "step": 2194
+    },
+    {
+      "epoch": 2.6932515337423313,
+      "grad_norm": 2.717301083812573,
+      "learning_rate": 8.264462809917355e-08,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -492.0,
+      "logps/rejected": -464.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.125,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -13.125,
+      "step": 2195
+    },
+    {
+      "epoch": 2.694478527607362,
+      "grad_norm": 1.2901660102619672,
+      "learning_rate": 8.231404958677686e-08,
+      "logits/chosen": -0.43359375,
+      "logits/rejected": -0.66796875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0029,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.07470703125,
+      "rewards/margins": 13.375,
+      "rewards/rejected": -13.3125,
+      "step": 2196
+    },
+    {
+      "epoch": 2.6957055214723926,
+      "grad_norm": 1.2475496828049488,
+      "learning_rate": 8.198347107438016e-08,
+      "logits/chosen": -0.33203125,
+      "logits/rejected": -0.62109375,
+      "logps/chosen": -448.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0023,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.37109375,
+      "rewards/margins": 14.0625,
+      "rewards/rejected": -13.6875,
+      "step": 2197
+    },
+    {
+      "epoch": 2.6969325153374233,
+      "grad_norm": 2.6414367845846627,
+      "learning_rate": 8.165289256198347e-08,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -406.0,
+      "loss": 0.005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.62890625,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.8125,
+      "step": 2198
+    },
+    {
+      "epoch": 2.698159509202454,
+      "grad_norm": 2.079060957920748,
+      "learning_rate": 8.132231404958677e-08,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -452.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.41015625,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.125,
+      "step": 2199
+    },
+    {
+      "epoch": 2.6993865030674846,
+      "grad_norm": 1.3019537401530699,
+      "learning_rate": 8.099173553719008e-08,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0038,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.51953125,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.625,
+      "step": 2200
+    },
+    {
+      "epoch": 2.7006134969325153,
+      "grad_norm": 2.3063150203343796,
+      "learning_rate": 8.066115702479338e-08,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.59765625,
+      "logps/chosen": -460.0,
+      "logps/rejected": -458.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 14.0,
+      "rewards/rejected": -12.9375,
+      "step": 2201
+    },
+    {
+      "epoch": 2.701840490797546,
+      "grad_norm": 8.925097244628772,
+      "learning_rate": 8.033057851239668e-08,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0085,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5703125,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.5625,
+      "step": 2202
+    },
+    {
+      "epoch": 2.7030674846625766,
+      "grad_norm": 2.8263734914079786,
+      "learning_rate": 8e-08,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.9296875,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.5625,
+      "step": 2203
+    },
+    {
+      "epoch": 2.7042944785276073,
+      "grad_norm": 1.1258416293468652,
+      "learning_rate": 7.96694214876033e-08,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0021,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.384765625,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -13.6875,
+      "step": 2204
+    },
+    {
+      "epoch": 2.705521472392638,
+      "grad_norm": 1.4377017298502133,
+      "learning_rate": 7.933884297520661e-08,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.68359375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.875,
+      "step": 2205
+    },
+    {
+      "epoch": 2.7067484662576686,
+      "grad_norm": 0.8838381979296313,
+      "learning_rate": 7.900826446280992e-08,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0076,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.64453125,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.375,
+      "step": 2206
+    },
+    {
+      "epoch": 2.7079754601226993,
+      "grad_norm": 2.082893707045185,
+      "learning_rate": 7.867768595041322e-08,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.83984375,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.75,
+      "step": 2207
+    },
+    {
+      "epoch": 2.70920245398773,
+      "grad_norm": 3.074242949317443,
+      "learning_rate": 7.834710743801652e-08,
+      "logits/chosen": -0.380859375,
+      "logits/rejected": -0.6328125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.134765625,
+      "rewards/margins": 13.875,
+      "rewards/rejected": -14.0,
+      "step": 2208
+    },
+    {
+      "epoch": 2.7104294478527606,
+      "grad_norm": 5.136148363448961,
+      "learning_rate": 7.801652892561983e-08,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0134,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.341796875,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -13.1875,
+      "step": 2209
+    },
+    {
+      "epoch": 2.7116564417177913,
+      "grad_norm": 4.403982620111939,
+      "learning_rate": 7.768595041322313e-08,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0086,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.734375,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.1875,
+      "step": 2210
+    },
+    {
+      "epoch": 2.712883435582822,
+      "grad_norm": 2.5557388930901785,
+      "learning_rate": 7.735537190082644e-08,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.15625,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -13.3125,
+      "step": 2211
+    },
+    {
+      "epoch": 2.7141104294478526,
+      "grad_norm": 4.519158707462654,
+      "learning_rate": 7.702479338842976e-08,
+      "logits/chosen": -0.35546875,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0126,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.83203125,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.25,
+      "step": 2212
+    },
+    {
+      "epoch": 2.7153374233128833,
+      "grad_norm": 2.56704044154945,
+      "learning_rate": 7.669421487603306e-08,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -432.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.275390625,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.0,
+      "step": 2213
+    },
+    {
+      "epoch": 2.716564417177914,
+      "grad_norm": 1.9833239617963656,
+      "learning_rate": 7.636363636363636e-08,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0058,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.69921875,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.875,
+      "step": 2214
+    },
+    {
+      "epoch": 2.7177914110429446,
+      "grad_norm": 2.5508126434970415,
+      "learning_rate": 7.603305785123967e-08,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.55859375,
+      "rewards/margins": 13.375,
+      "rewards/rejected": -13.9375,
+      "step": 2215
+    },
+    {
+      "epoch": 2.7190184049079753,
+      "grad_norm": 15.617386469652637,
+      "learning_rate": 7.570247933884297e-08,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0176,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.84375,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.125,
+      "step": 2216
+    },
+    {
+      "epoch": 2.720245398773006,
+      "grad_norm": 1.3210369359927996,
+      "learning_rate": 7.537190082644628e-08,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0081,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.041259765625,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -12.6875,
+      "step": 2217
+    },
+    {
+      "epoch": 2.721472392638037,
+      "grad_norm": 2.1940261882390315,
+      "learning_rate": 7.504132231404958e-08,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0101,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.60546875,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -13.5,
+      "step": 2218
+    },
+    {
+      "epoch": 2.7226993865030673,
+      "grad_norm": 1.8304595035834146,
+      "learning_rate": 7.471074380165289e-08,
+      "logits/chosen": -0.248046875,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.57421875,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.25,
+      "step": 2219
+    },
+    {
+      "epoch": 2.7239263803680984,
+      "grad_norm": 2.3966944940327592,
+      "learning_rate": 7.438016528925619e-08,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -442.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0077,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.58203125,
+      "rewards/margins": 13.5,
+      "rewards/rejected": -12.875,
+      "step": 2220
+    },
+    {
+      "epoch": 2.7251533742331286,
+      "grad_norm": 1.13146060739838,
+      "learning_rate": 7.40495867768595e-08,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0111,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.59765625,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -13.75,
+      "step": 2221
+    },
+    {
+      "epoch": 2.7263803680981598,
+      "grad_norm": 1.5353375263726436,
+      "learning_rate": 7.371900826446281e-08,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.90625,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -13.25,
+      "step": 2222
+    },
+    {
+      "epoch": 2.72760736196319,
+      "grad_norm": 1.6975022176181414,
+      "learning_rate": 7.338842975206612e-08,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2431640625,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -13.375,
+      "step": 2223
+    },
+    {
+      "epoch": 2.728834355828221,
+      "grad_norm": 10.090032983305967,
+      "learning_rate": 7.305785123966942e-08,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0172,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.1875,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.875,
+      "step": 2224
+    },
+    {
+      "epoch": 2.7300613496932513,
+      "grad_norm": 2.446424279347937,
+      "learning_rate": 7.272727272727273e-08,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0103,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.6875,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.875,
+      "step": 2225
+    },
+    {
+      "epoch": 2.7312883435582824,
+      "grad_norm": 1.499316518750317,
+      "learning_rate": 7.239669421487603e-08,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -448.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.609375,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.3125,
+      "step": 2226
+    },
+    {
+      "epoch": 2.7325153374233127,
+      "grad_norm": 1.5221373016779374,
+      "learning_rate": 7.206611570247934e-08,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -350.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1015625,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -12.6875,
+      "step": 2227
+    },
+    {
+      "epoch": 2.7337423312883438,
+      "grad_norm": 6.572307889036707,
+      "learning_rate": 7.173553719008264e-08,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0066,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.57421875,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -13.5,
+      "step": 2228
+    },
+    {
+      "epoch": 2.734969325153374,
+      "grad_norm": 1.5333165216603437,
+      "learning_rate": 7.140495867768594e-08,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0830078125,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.625,
+      "step": 2229
+    },
+    {
+      "epoch": 2.736196319018405,
+      "grad_norm": 8.701937133897674,
+      "learning_rate": 7.107438016528925e-08,
+      "logits/chosen": -0.41015625,
+      "logits/rejected": -0.625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0079,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.09228515625,
+      "rewards/margins": 14.0,
+      "rewards/rejected": -13.9375,
+      "step": 2230
+    },
+    {
+      "epoch": 2.7374233128834353,
+      "grad_norm": 2.4587809793552493,
+      "learning_rate": 7.074380165289255e-08,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.609375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.10107421875,
+      "rewards/margins": 13.5,
+      "rewards/rejected": -13.625,
+      "step": 2231
+    },
+    {
+      "epoch": 2.7386503067484664,
+      "grad_norm": 2.39191191911374,
+      "learning_rate": 7.041322314049587e-08,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.265625,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.5625,
+      "step": 2232
+    },
+    {
+      "epoch": 2.7398773006134967,
+      "grad_norm": 2.5490132308128737,
+      "learning_rate": 7.008264462809918e-08,
+      "logits/chosen": -0.380859375,
+      "logits/rejected": -0.61328125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -442.0,
+      "loss": 0.0058,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.404296875,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -13.3125,
+      "step": 2233
+    },
+    {
+      "epoch": 2.7411042944785278,
+      "grad_norm": 1.0506484012932482,
+      "learning_rate": 6.975206611570248e-08,
+      "logits/chosen": -0.376953125,
+      "logits/rejected": -0.60546875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -442.0,
+      "loss": 0.0081,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.451171875,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -13.75,
+      "step": 2234
+    },
+    {
+      "epoch": 2.7423312883435584,
+      "grad_norm": 1.4483748839658306,
+      "learning_rate": 6.942148760330578e-08,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -332.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0089,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.73828125,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.5,
+      "step": 2235
+    },
+    {
+      "epoch": 2.743558282208589,
+      "grad_norm": 1.5257272713174417,
+      "learning_rate": 6.909090909090909e-08,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -448.0,
+      "logps/rejected": -474.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.162109375,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.4375,
+      "step": 2236
+    },
+    {
+      "epoch": 2.7447852760736198,
+      "grad_norm": 1.697954391009891,
+      "learning_rate": 6.876033057851239e-08,
+      "logits/chosen": -0.337890625,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0057,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6953125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.0,
+      "step": 2237
+    },
+    {
+      "epoch": 2.7460122699386504,
+      "grad_norm": 2.0900347983642447,
+      "learning_rate": 6.84297520661157e-08,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.86328125,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -13.875,
+      "step": 2238
+    },
+    {
+      "epoch": 2.747239263803681,
+      "grad_norm": 2.0021189450767345,
+      "learning_rate": 6.8099173553719e-08,
+      "logits/chosen": -0.341796875,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0091,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.37109375,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.0625,
+      "step": 2239
+    },
+    {
+      "epoch": 2.7484662576687118,
+      "grad_norm": 1.1985328989941613,
+      "learning_rate": 6.77685950413223e-08,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -430.0,
+      "loss": 0.014,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.478515625,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.0,
+      "step": 2240
+    },
+    {
+      "epoch": 2.7496932515337424,
+      "grad_norm": 5.603696138773228,
+      "learning_rate": 6.743801652892561e-08,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0075,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0625,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -13.0625,
+      "step": 2241
+    },
+    {
+      "epoch": 2.750920245398773,
+      "grad_norm": 2.484784063813252,
+      "learning_rate": 6.710743801652893e-08,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -352.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0057,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.478515625,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.625,
+      "step": 2242
+    },
+    {
+      "epoch": 2.752147239263804,
+      "grad_norm": 1.8099869310925514,
+      "learning_rate": 6.677685950413223e-08,
+      "logits/chosen": -0.3515625,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0088,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.796875,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.5625,
+      "step": 2243
+    },
+    {
+      "epoch": 2.7533742331288344,
+      "grad_norm": 2.599969430923648,
+      "learning_rate": 6.644628099173554e-08,
+      "logits/chosen": -0.357421875,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -398.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0112,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.1953125,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.875,
+      "step": 2244
+    },
+    {
+      "epoch": 2.754601226993865,
+      "grad_norm": 1.8959409680007666,
+      "learning_rate": 6.611570247933884e-08,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.091796875,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -12.5625,
+      "step": 2245
+    },
+    {
+      "epoch": 2.755828220858896,
+      "grad_norm": 2.0375129943781434,
+      "learning_rate": 6.578512396694215e-08,
+      "logits/chosen": -0.365234375,
+      "logits/rejected": -0.57421875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8984375,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -13.0,
+      "step": 2246
+    },
+    {
+      "epoch": 2.7570552147239265,
+      "grad_norm": 2.2942599267172206,
+      "learning_rate": 6.545454545454545e-08,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -450.0,
+      "logps/rejected": -442.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.08984375,
+      "rewards/margins": 14.125,
+      "rewards/rejected": -14.25,
+      "step": 2247
+    },
+    {
+      "epoch": 2.758282208588957,
+      "grad_norm": 2.952119004696011,
+      "learning_rate": 6.512396694214876e-08,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1171875,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -13.125,
+      "step": 2248
+    },
+    {
+      "epoch": 2.759509202453988,
+      "grad_norm": 21.129165007315503,
+      "learning_rate": 6.479338842975206e-08,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -386.0,
+      "loss": 0.019,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.1669921875,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -11.8125,
+      "step": 2249
+    },
+    {
+      "epoch": 2.7607361963190185,
+      "grad_norm": 2.1702702024986644,
+      "learning_rate": 6.446280991735536e-08,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.61328125,
+      "rewards/margins": 11.1875,
+      "rewards/rejected": -11.8125,
+      "step": 2250
+    },
+    {
+      "epoch": 2.761963190184049,
+      "grad_norm": 2.6722133425770624,
+      "learning_rate": 6.413223140495867e-08,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -392.0,
+      "loss": 0.006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0703125,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -12.4375,
+      "step": 2251
+    },
+    {
+      "epoch": 2.76319018404908,
+      "grad_norm": 2.539703403154541,
+      "learning_rate": 6.380165289256199e-08,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -464.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.94140625,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -13.4375,
+      "step": 2252
+    },
+    {
+      "epoch": 2.7644171779141105,
+      "grad_norm": 2.2587503435485474,
+      "learning_rate": 6.347107438016529e-08,
+      "logits/chosen": -0.404296875,
+      "logits/rejected": -0.55078125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0038,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.2265625,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -13.375,
+      "step": 2253
+    },
+    {
+      "epoch": 2.765644171779141,
+      "grad_norm": 2.6179757111415594,
+      "learning_rate": 6.31404958677686e-08,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0114,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.08349609375,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.625,
+      "step": 2254
+    },
+    {
+      "epoch": 2.766871165644172,
+      "grad_norm": 2.620995241226507,
+      "learning_rate": 6.28099173553719e-08,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.609375,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.3125,
+      "step": 2255
+    },
+    {
+      "epoch": 2.7680981595092025,
+      "grad_norm": 3.545197005274326,
+      "learning_rate": 6.24793388429752e-08,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0085,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0625,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -13.1875,
+      "step": 2256
+    },
+    {
+      "epoch": 2.769325153374233,
+      "grad_norm": 1.0292424246825402,
+      "learning_rate": 6.214876033057851e-08,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -444.0,
+      "loss": 0.0022,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8984375,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.5,
+      "step": 2257
+    },
+    {
+      "epoch": 2.770552147239264,
+      "grad_norm": 1.3943670382796598,
+      "learning_rate": 6.181818181818181e-08,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8671875,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.4375,
+      "step": 2258
+    },
+    {
+      "epoch": 2.7717791411042945,
+      "grad_norm": 2.542940036826831,
+      "learning_rate": 6.148760330578512e-08,
+      "logits/chosen": -0.357421875,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -440.0,
+      "logps/rejected": -450.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.79296875,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.8125,
+      "step": 2259
+    },
+    {
+      "epoch": 2.773006134969325,
+      "grad_norm": 1.765448796157598,
+      "learning_rate": 6.115702479338842e-08,
+      "logits/chosen": -0.396484375,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.34375,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -13.875,
+      "step": 2260
+    },
+    {
+      "epoch": 2.774233128834356,
+      "grad_norm": 5.557935848950858,
+      "learning_rate": 6.082644628099173e-08,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -448.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.31640625,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.5625,
+      "step": 2261
+    },
+    {
+      "epoch": 2.7754601226993865,
+      "grad_norm": 1.9056186049004322,
+      "learning_rate": 6.049586776859504e-08,
+      "logits/chosen": -0.39453125,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -440.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0036,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.671875,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -13.75,
+      "step": 2262
+    },
+    {
+      "epoch": 2.776687116564417,
+      "grad_norm": 0.6031421959929165,
+      "learning_rate": 6.016528925619835e-08,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0068,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.72265625,
+      "rewards/margins": 13.625,
+      "rewards/rejected": -14.3125,
+      "step": 2263
+    },
+    {
+      "epoch": 2.777914110429448,
+      "grad_norm": 1.8470774442522424,
+      "learning_rate": 5.983471074380165e-08,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.97265625,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -13.125,
+      "step": 2264
+    },
+    {
+      "epoch": 2.7791411042944785,
+      "grad_norm": 1.5695995832322116,
+      "learning_rate": 5.950413223140496e-08,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.40234375,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.0,
+      "step": 2265
+    },
+    {
+      "epoch": 2.780368098159509,
+      "grad_norm": 1.8303521007764108,
+      "learning_rate": 5.917355371900826e-08,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.54296875,
+      "logps/chosen": -452.0,
+      "logps/rejected": -424.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.29296875,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -13.375,
+      "step": 2266
+    },
+    {
+      "epoch": 2.78159509202454,
+      "grad_norm": 1.6434841062190646,
+      "learning_rate": 5.884297520661157e-08,
+      "logits/chosen": -0.359375,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -476.0,
+      "logps/rejected": -440.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2021484375,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.0625,
+      "step": 2267
+    },
+    {
+      "epoch": 2.7828220858895705,
+      "grad_norm": 1.4755906375390189,
+      "learning_rate": 5.851239669421487e-08,
+      "logits/chosen": -0.1201171875,
+      "logits/rejected": -0.234375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5078125,
+      "rewards/margins": 11.5625,
+      "rewards/rejected": -12.0625,
+      "step": 2268
+    },
+    {
+      "epoch": 2.784049079754601,
+      "grad_norm": 7.511191109853007,
+      "learning_rate": 5.8181818181818176e-08,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.490234375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -406.0,
+      "loss": 0.01,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.236328125,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.75,
+      "step": 2269
+    },
+    {
+      "epoch": 2.785276073619632,
+      "grad_norm": 1.848729290975255,
+      "learning_rate": 5.785123966942148e-08,
+      "logits/chosen": -0.41015625,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.96875,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -13.3125,
+      "step": 2270
+    },
+    {
+      "epoch": 2.7865030674846625,
+      "grad_norm": 1.5655467769953058,
+      "learning_rate": 5.7520661157024785e-08,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0091,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.1015625,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.9375,
+      "step": 2271
+    },
+    {
+      "epoch": 2.787730061349693,
+      "grad_norm": 2.368097354925877,
+      "learning_rate": 5.71900826446281e-08,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0103,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.52734375,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.25,
+      "step": 2272
+    },
+    {
+      "epoch": 2.788957055214724,
+      "grad_norm": 1.996192598307064,
+      "learning_rate": 5.685950413223141e-08,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -454.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.12890625,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.9375,
+      "step": 2273
+    },
+    {
+      "epoch": 2.7901840490797545,
+      "grad_norm": 2.108965446470985,
+      "learning_rate": 5.652892561983471e-08,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.5,
+      "logps/chosen": -442.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.138671875,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.1875,
+      "step": 2274
+    },
+    {
+      "epoch": 2.791411042944785,
+      "grad_norm": 2.37754205005712,
+      "learning_rate": 5.6198347107438016e-08,
+      "logits/chosen": -0.380859375,
+      "logits/rejected": -0.58203125,
+      "logps/chosen": -454.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.828125,
+      "rewards/margins": 13.5625,
+      "rewards/rejected": -12.75,
+      "step": 2275
+    },
+    {
+      "epoch": 2.792638036809816,
+      "grad_norm": 4.900314435907703,
+      "learning_rate": 5.586776859504132e-08,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0136,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.1767578125,
+      "rewards/margins": 13.375,
+      "rewards/rejected": -13.25,
+      "step": 2276
+    },
+    {
+      "epoch": 2.7938650306748465,
+      "grad_norm": 2.2539666890493044,
+      "learning_rate": 5.5537190082644625e-08,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.265625,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.8125,
+      "step": 2277
+    },
+    {
+      "epoch": 2.795092024539877,
+      "grad_norm": 1.7135731505882545,
+      "learning_rate": 5.520661157024793e-08,
+      "logits/chosen": -0.2890625,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0038,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.10791015625,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -13.1875,
+      "step": 2278
+    },
+    {
+      "epoch": 2.796319018404908,
+      "grad_norm": 1.2200184045571503,
+      "learning_rate": 5.4876033057851234e-08,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -406.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.048828125,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.9375,
+      "step": 2279
+    },
+    {
+      "epoch": 2.7975460122699385,
+      "grad_norm": 3.807953561772992,
+      "learning_rate": 5.454545454545454e-08,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -456.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0063,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.392578125,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.9375,
+      "step": 2280
+    },
+    {
+      "epoch": 2.7987730061349696,
+      "grad_norm": 2.014829441602722,
+      "learning_rate": 5.421487603305784e-08,
+      "logits/chosen": -0.369140625,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -460.0,
+      "logps/rejected": -450.0,
+      "loss": 0.0038,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1708984375,
+      "rewards/margins": 13.625,
+      "rewards/rejected": -13.75,
+      "step": 2281
+    },
+    {
+      "epoch": 2.8,
+      "grad_norm": 1.7735442585245964,
+      "learning_rate": 5.388429752066116e-08,
+      "logits/chosen": -0.33984375,
+      "logits/rejected": -0.58203125,
+      "logps/chosen": -492.0,
+      "logps/rejected": -460.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2294921875,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -13.0625,
+      "step": 2282
+    },
+    {
+      "epoch": 2.801226993865031,
+      "grad_norm": 1.5001532112578835,
+      "learning_rate": 5.3553719008264465e-08,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.439453125,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -13.0,
+      "step": 2283
+    },
+    {
+      "epoch": 2.802453987730061,
+      "grad_norm": 1.0976052882314502,
+      "learning_rate": 5.322314049586777e-08,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0023,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.10302734375,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.9375,
+      "step": 2284
+    },
+    {
+      "epoch": 2.8036809815950923,
+      "grad_norm": 3.5334538925832284,
+      "learning_rate": 5.2892561983471074e-08,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0058,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.390625,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.875,
+      "step": 2285
+    },
+    {
+      "epoch": 2.8049079754601225,
+      "grad_norm": 1.8967992373881246,
+      "learning_rate": 5.256198347107438e-08,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -448.0,
+      "loss": 0.0029,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.02001953125,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.75,
+      "step": 2286
+    },
+    {
+      "epoch": 2.8061349693251536,
+      "grad_norm": 4.625657713530512,
+      "learning_rate": 5.223140495867768e-08,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.003814697265625,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.0625,
+      "step": 2287
+    },
+    {
+      "epoch": 2.807361963190184,
+      "grad_norm": 1.3364916671459077,
+      "learning_rate": 5.190082644628099e-08,
+      "logits/chosen": -0.375,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0087,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.86328125,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.4375,
+      "step": 2288
+    },
+    {
+      "epoch": 2.808588957055215,
+      "grad_norm": 3.154704964575447,
+      "learning_rate": 5.157024793388429e-08,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -428.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0121,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.240234375,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -13.125,
+      "step": 2289
+    },
+    {
+      "epoch": 2.809815950920245,
+      "grad_norm": 1.8033458126551494,
+      "learning_rate": 5.1239669421487596e-08,
+      "logits/chosen": -0.29296875,
+      "logits/rejected": -0.49609375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.4296875,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.0,
+      "step": 2290
+    },
+    {
+      "epoch": 2.8110429447852763,
+      "grad_norm": 10.689735615602745,
+      "learning_rate": 5.09090909090909e-08,
+      "logits/chosen": -0.341796875,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0406,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.2275390625,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -12.9375,
+      "step": 2291
+    },
+    {
+      "epoch": 2.8122699386503065,
+      "grad_norm": 2.320187720592824,
+      "learning_rate": 5.057851239669422e-08,
+      "logits/chosen": -0.38671875,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.88671875,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -13.9375,
+      "step": 2292
+    },
+    {
+      "epoch": 2.8134969325153376,
+      "grad_norm": 1.623467891985287,
+      "learning_rate": 5.024793388429752e-08,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -516.0,
+      "logps/rejected": -462.0,
+      "loss": 0.0083,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.92578125,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -12.0625,
+      "step": 2293
+    },
+    {
+      "epoch": 2.814723926380368,
+      "grad_norm": 1.4990940263623278,
+      "learning_rate": 4.991735537190083e-08,
+      "logits/chosen": -0.408203125,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0084,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.390625,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -13.3125,
+      "step": 2294
+    },
+    {
+      "epoch": 2.815950920245399,
+      "grad_norm": 1.2765494027640394,
+      "learning_rate": 4.958677685950413e-08,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0083,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.5859375,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -13.375,
+      "step": 2295
+    },
+    {
+      "epoch": 2.817177914110429,
+      "grad_norm": 8.441906480997861,
+      "learning_rate": 4.9256198347107436e-08,
+      "logits/chosen": -0.3671875,
+      "logits/rejected": -0.59375,
+      "logps/chosen": -450.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0095,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.107421875,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -12.4375,
+      "step": 2296
+    },
+    {
+      "epoch": 2.8184049079754603,
+      "grad_norm": 2.295304145533487,
+      "learning_rate": 4.892561983471074e-08,
+      "logits/chosen": -0.365234375,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0045,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.3203125,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -13.4375,
+      "step": 2297
+    },
+    {
+      "epoch": 2.819631901840491,
+      "grad_norm": 2.424740840515899,
+      "learning_rate": 4.8595041322314045e-08,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -352.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0108,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.349609375,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.25,
+      "step": 2298
+    },
+    {
+      "epoch": 2.8208588957055216,
+      "grad_norm": 2.8622521983495157,
+      "learning_rate": 4.8264462809917356e-08,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0083,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6015625,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.4375,
+      "step": 2299
+    },
+    {
+      "epoch": 2.8220858895705523,
+      "grad_norm": 2.452648418396362,
+      "learning_rate": 4.793388429752066e-08,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -438.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.4765625,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -13.6875,
+      "step": 2300
+    },
+    {
+      "epoch": 2.823312883435583,
+      "grad_norm": 2.319620976722443,
+      "learning_rate": 4.7603305785123965e-08,
+      "logits/chosen": -0.3671875,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -352.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.458984375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.6875,
+      "step": 2301
+    },
+    {
+      "epoch": 2.8245398773006136,
+      "grad_norm": 3.7298541958736715,
+      "learning_rate": 4.727272727272727e-08,
+      "logits/chosen": -0.46875,
+      "logits/rejected": -0.67578125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2734375,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.0625,
+      "step": 2302
+    },
+    {
+      "epoch": 2.8257668711656443,
+      "grad_norm": 3.5761828040496524,
+      "learning_rate": 4.6942148760330574e-08,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.77734375,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -12.1875,
+      "step": 2303
+    },
+    {
+      "epoch": 2.826993865030675,
+      "grad_norm": 3.4770327041687414,
+      "learning_rate": 4.6611570247933885e-08,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -396.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.328125,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.375,
+      "step": 2304
+    },
+    {
+      "epoch": 2.8282208588957056,
+      "grad_norm": 1.8544951083588206,
+      "learning_rate": 4.628099173553719e-08,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -314.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0068,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.296875,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.75,
+      "step": 2305
+    },
+    {
+      "epoch": 2.8294478527607363,
+      "grad_norm": 1.0816575276582756,
+      "learning_rate": 4.5950413223140494e-08,
+      "logits/chosen": -0.37109375,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0087,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.0859375,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.9375,
+      "step": 2306
+    },
+    {
+      "epoch": 2.830674846625767,
+      "grad_norm": 2.5360053493219588,
+      "learning_rate": 4.56198347107438e-08,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0105,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.123046875,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.125,
+      "step": 2307
+    },
+    {
+      "epoch": 2.8319018404907976,
+      "grad_norm": 1.8457405395498343,
+      "learning_rate": 4.52892561983471e-08,
+      "logits/chosen": -0.400390625,
+      "logits/rejected": -0.60546875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -432.0,
+      "loss": 0.005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.38671875,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.625,
+      "step": 2308
+    },
+    {
+      "epoch": 2.8331288343558283,
+      "grad_norm": 3.1892046068455078,
+      "learning_rate": 4.4958677685950414e-08,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0093,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.49609375,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.8125,
+      "step": 2309
+    },
+    {
+      "epoch": 2.834355828220859,
+      "grad_norm": 3.078722560991968,
+      "learning_rate": 4.462809917355372e-08,
+      "logits/chosen": -0.365234375,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0056,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.396484375,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.125,
+      "step": 2310
+    },
+    {
+      "epoch": 2.8355828220858896,
+      "grad_norm": 0.7840969439625597,
+      "learning_rate": 4.429752066115702e-08,
+      "logits/chosen": -0.3984375,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0019,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.0062255859375,
+      "rewards/margins": 13.4375,
+      "rewards/rejected": -13.4375,
+      "step": 2311
+    },
+    {
+      "epoch": 2.8368098159509203,
+      "grad_norm": 3.6134505333349405,
+      "learning_rate": 4.396694214876033e-08,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2255859375,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -13.375,
+      "step": 2312
+    },
+    {
+      "epoch": 2.838036809815951,
+      "grad_norm": 5.032829862411979,
+      "learning_rate": 4.363636363636363e-08,
+      "logits/chosen": -0.373046875,
+      "logits/rejected": -0.59765625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0078,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.76171875,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.5,
+      "step": 2313
+    },
+    {
+      "epoch": 2.8392638036809816,
+      "grad_norm": 1.085357037720828,
+      "learning_rate": 4.330578512396694e-08,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.5,
+      "logps/chosen": -374.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1865234375,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -12.9375,
+      "step": 2314
+    },
+    {
+      "epoch": 2.8404907975460123,
+      "grad_norm": 3.123544866687054,
+      "learning_rate": 4.297520661157025e-08,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.470703125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.11474609375,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -13.125,
+      "step": 2315
+    },
+    {
+      "epoch": 2.841717791411043,
+      "grad_norm": 1.9628598021603092,
+      "learning_rate": 4.264462809917355e-08,
+      "logits/chosen": -0.44921875,
+      "logits/rejected": -0.6328125,
+      "logps/chosen": -332.0,
+      "logps/rejected": -372.0,
+      "loss": 0.011,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.6484375,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.375,
+      "step": 2316
+    },
+    {
+      "epoch": 2.8429447852760736,
+      "grad_norm": 2.398483232612997,
+      "learning_rate": 4.2314049586776856e-08,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0077,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6875,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -13.1875,
+      "step": 2317
+    },
+    {
+      "epoch": 2.8441717791411043,
+      "grad_norm": 2.9409677648355785,
+      "learning_rate": 4.198347107438016e-08,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.478515625,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.625,
+      "step": 2318
+    },
+    {
+      "epoch": 2.845398773006135,
+      "grad_norm": 3.4099778750574954,
+      "learning_rate": 4.165289256198347e-08,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0102,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.205078125,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -12.4375,
+      "step": 2319
+    },
+    {
+      "epoch": 2.8466257668711656,
+      "grad_norm": 2.953835843910011,
+      "learning_rate": 4.1322314049586776e-08,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0056,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2080078125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.5,
+      "step": 2320
+    },
+    {
+      "epoch": 2.8478527607361963,
+      "grad_norm": 3.4118181284623734,
+      "learning_rate": 4.099173553719008e-08,
+      "logits/chosen": -0.3828125,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0057,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2197265625,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.5625,
+      "step": 2321
+    },
+    {
+      "epoch": 2.849079754601227,
+      "grad_norm": 1.9295108354040367,
+      "learning_rate": 4.0661157024793385e-08,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.026611328125,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.5625,
+      "step": 2322
+    },
+    {
+      "epoch": 2.8503067484662576,
+      "grad_norm": 6.680191261748946,
+      "learning_rate": 4.033057851239669e-08,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0077,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5859375,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.25,
+      "step": 2323
+    },
+    {
+      "epoch": 2.8515337423312883,
+      "grad_norm": 2.3123044142818507,
+      "learning_rate": 4e-08,
+      "logits/chosen": -0.333984375,
+      "logits/rejected": -0.47265625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.63671875,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.5625,
+      "step": 2324
+    },
+    {
+      "epoch": 2.852760736196319,
+      "grad_norm": 1.9102237457542082,
+      "learning_rate": 3.9669421487603305e-08,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0144,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.03125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.3125,
+      "step": 2325
+    },
+    {
+      "epoch": 2.8539877300613496,
+      "grad_norm": 1.7408309272472986,
+      "learning_rate": 3.933884297520661e-08,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.490234375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.28125,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -13.0,
+      "step": 2326
+    },
+    {
+      "epoch": 2.8552147239263803,
+      "grad_norm": 1.1478187146635819,
+      "learning_rate": 3.9008264462809914e-08,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -404.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.67578125,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.5,
+      "step": 2327
+    },
+    {
+      "epoch": 2.856441717791411,
+      "grad_norm": 1.4521227108447212,
+      "learning_rate": 3.867768595041322e-08,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.06787109375,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.1875,
+      "step": 2328
+    },
+    {
+      "epoch": 2.8576687116564417,
+      "grad_norm": 3.531696397099087,
+      "learning_rate": 3.834710743801653e-08,
+      "logits/chosen": -0.38671875,
+      "logits/rejected": -0.61328125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0079,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.4296875,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -14.25,
+      "step": 2329
+    },
+    {
+      "epoch": 2.8588957055214723,
+      "grad_norm": 8.486941746767686,
+      "learning_rate": 3.8016528925619834e-08,
+      "logits/chosen": -0.322265625,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -452.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0107,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.318359375,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.9375,
+      "step": 2330
+    },
+    {
+      "epoch": 2.860122699386503,
+      "grad_norm": 1.1747245219364273,
+      "learning_rate": 3.768595041322314e-08,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.5,
+      "logps/chosen": -398.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0028,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.4375,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -13.25,
+      "step": 2331
+    },
+    {
+      "epoch": 2.8613496932515337,
+      "grad_norm": 2.1800113233442344,
+      "learning_rate": 3.735537190082644e-08,
+      "logits/chosen": -0.34765625,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.173828125,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.0625,
+      "step": 2332
+    },
+    {
+      "epoch": 2.8625766871165643,
+      "grad_norm": 1.255711370439629,
+      "learning_rate": 3.702479338842975e-08,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -416.0,
+      "loss": 0.0029,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.396484375,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.0625,
+      "step": 2333
+    },
+    {
+      "epoch": 2.863803680981595,
+      "grad_norm": 2.2596508726142837,
+      "learning_rate": 3.669421487603306e-08,
+      "logits/chosen": -0.37109375,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0052,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1328125,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -13.125,
+      "step": 2334
+    },
+    {
+      "epoch": 2.8650306748466257,
+      "grad_norm": 1.5584375660034178,
+      "learning_rate": 3.636363636363636e-08,
+      "logits/chosen": -0.431640625,
+      "logits/rejected": -0.5703125,
+      "logps/chosen": -454.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0038,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.455078125,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -13.5625,
+      "step": 2335
+    },
+    {
+      "epoch": 2.8662576687116563,
+      "grad_norm": 5.331509541013691,
+      "learning_rate": 3.603305785123967e-08,
+      "logits/chosen": -0.439453125,
+      "logits/rejected": -0.625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.51953125,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.125,
+      "step": 2336
+    },
+    {
+      "epoch": 2.867484662576687,
+      "grad_norm": 0.9086085307479842,
+      "learning_rate": 3.570247933884297e-08,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.7421875,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.875,
+      "step": 2337
+    },
+    {
+      "epoch": 2.8687116564417177,
+      "grad_norm": 2.752850636176483,
+      "learning_rate": 3.5371900826446277e-08,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0079,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.125,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -12.4375,
+      "step": 2338
+    },
+    {
+      "epoch": 2.8699386503067483,
+      "grad_norm": 1.22792645912641,
+      "learning_rate": 3.504132231404959e-08,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.05126953125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.3125,
+      "step": 2339
+    },
+    {
+      "epoch": 2.871165644171779,
+      "grad_norm": 1.3509568578363804,
+      "learning_rate": 3.471074380165289e-08,
+      "logits/chosen": -0.330078125,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -428.0,
+      "logps/rejected": -446.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.9296875,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.625,
+      "step": 2340
+    },
+    {
+      "epoch": 2.8723926380368097,
+      "grad_norm": 1.2268444589186553,
+      "learning_rate": 3.4380165289256197e-08,
+      "logits/chosen": -0.365234375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -368.0,
+      "logps/rejected": -432.0,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.107421875,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.8125,
+      "step": 2341
+    },
+    {
+      "epoch": 2.8736196319018403,
+      "grad_norm": 3.415967256742802,
+      "learning_rate": 3.40495867768595e-08,
+      "logits/chosen": -0.44921875,
+      "logits/rejected": -0.6953125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -432.0,
+      "loss": 0.0074,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.17578125,
+      "rewards/margins": 13.5625,
+      "rewards/rejected": -13.6875,
+      "step": 2342
+    },
+    {
+      "epoch": 2.874846625766871,
+      "grad_norm": 1.1270164719135747,
+      "learning_rate": 3.3719008264462806e-08,
+      "logits/chosen": -0.373046875,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0028,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.49609375,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.625,
+      "step": 2343
+    },
+    {
+      "epoch": 2.876073619631902,
+      "grad_norm": 5.809614326223801,
+      "learning_rate": 3.3388429752066117e-08,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0057,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.28515625,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -13.125,
+      "step": 2344
+    },
+    {
+      "epoch": 2.8773006134969323,
+      "grad_norm": 1.3412149874370591,
+      "learning_rate": 3.305785123966942e-08,
+      "logits/chosen": -0.4140625,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0024,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2099609375,
+      "rewards/margins": 13.1875,
+      "rewards/rejected": -13.4375,
+      "step": 2345
+    },
+    {
+      "epoch": 2.8785276073619634,
+      "grad_norm": 2.4232253905956305,
+      "learning_rate": 3.2727272727272726e-08,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -456.0,
+      "logps/rejected": -432.0,
+      "loss": 0.0092,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.2236328125,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.0,
+      "step": 2346
+    },
+    {
+      "epoch": 2.8797546012269937,
+      "grad_norm": 2.4622563175390773,
+      "learning_rate": 3.239669421487603e-08,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0067,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.62109375,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -13.5625,
+      "step": 2347
+    },
+    {
+      "epoch": 2.880981595092025,
+      "grad_norm": 3.554112846729448,
+      "learning_rate": 3.2066115702479334e-08,
+      "logits/chosen": -0.2392578125,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -398.0,
+      "loss": 0.006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.369140625,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.6875,
+      "step": 2348
+    },
+    {
+      "epoch": 2.882208588957055,
+      "grad_norm": 1.3641749976906659,
+      "learning_rate": 3.1735537190082646e-08,
+      "logits/chosen": -0.3671875,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -392.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0390625,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -13.375,
+      "step": 2349
+    },
+    {
+      "epoch": 2.883435582822086,
+      "grad_norm": 1.34395588414787,
+      "learning_rate": 3.140495867768595e-08,
+      "logits/chosen": -0.419921875,
+      "logits/rejected": -0.6796875,
+      "logps/chosen": -452.0,
+      "logps/rejected": -446.0,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.68359375,
+      "rewards/margins": 13.8125,
+      "rewards/rejected": -13.125,
+      "step": 2350
+    },
+    {
+      "epoch": 2.8846625766871163,
+      "grad_norm": 1.450992640582867,
+      "learning_rate": 3.1074380165289254e-08,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0201,
+      "rewards/accuracies": 0.9765625,
+      "rewards/chosen": -0.96484375,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.1875,
+      "step": 2351
+    },
+    {
+      "epoch": 2.8858895705521475,
+      "grad_norm": 2.2106350912526653,
+      "learning_rate": 3.074380165289256e-08,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.671875,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.25,
+      "step": 2352
+    },
+    {
+      "epoch": 2.8871165644171777,
+      "grad_norm": 1.128180709653168,
+      "learning_rate": 3.0413223140495863e-08,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2392578125,
+      "rewards/margins": 13.5,
+      "rewards/rejected": -13.75,
+      "step": 2353
+    },
+    {
+      "epoch": 2.888343558282209,
+      "grad_norm": 0.8613807004777035,
+      "learning_rate": 3.0082644628099175e-08,
+      "logits/chosen": -0.390625,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0022,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.50390625,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.5625,
+      "step": 2354
+    },
+    {
+      "epoch": 2.889570552147239,
+      "grad_norm": 3.6845356009274526,
+      "learning_rate": 2.975206611570248e-08,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0062,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5703125,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.375,
+      "step": 2355
+    },
+    {
+      "epoch": 2.89079754601227,
+      "grad_norm": 1.4347823915578544,
+      "learning_rate": 2.9421487603305783e-08,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0028,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.01141357421875,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -12.6875,
+      "step": 2356
+    },
+    {
+      "epoch": 2.8920245398773003,
+      "grad_norm": 2.1028420785679707,
+      "learning_rate": 2.9090909090909088e-08,
+      "logits/chosen": -0.48046875,
+      "logits/rejected": -0.59765625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.81640625,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -13.125,
+      "step": 2357
+    },
+    {
+      "epoch": 2.8932515337423315,
+      "grad_norm": 2.511394661485255,
+      "learning_rate": 2.8760330578512392e-08,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.034423828125,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.375,
+      "step": 2358
+    },
+    {
+      "epoch": 2.894478527607362,
+      "grad_norm": 1.9853598293094303,
+      "learning_rate": 2.8429752066115703e-08,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -482.0,
+      "logps/rejected": -456.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1748046875,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.4375,
+      "step": 2359
+    },
+    {
+      "epoch": 2.895705521472393,
+      "grad_norm": 22.77327017062875,
+      "learning_rate": 2.8099173553719008e-08,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -390.0,
+      "loss": 0.0214,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": -0.77734375,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.5,
+      "step": 2360
+    },
+    {
+      "epoch": 2.8969325153374235,
+      "grad_norm": 2.6943557267688845,
+      "learning_rate": 2.7768595041322312e-08,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.04248046875,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -12.9375,
+      "step": 2361
+    },
+    {
+      "epoch": 2.898159509202454,
+      "grad_norm": 1.2565960476608073,
+      "learning_rate": 2.7438016528925617e-08,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0024,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.059326171875,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.875,
+      "step": 2362
+    },
+    {
+      "epoch": 2.899386503067485,
+      "grad_norm": 1.9017508396109692,
+      "learning_rate": 2.710743801652892e-08,
+      "logits/chosen": -0.41015625,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.46484375,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -13.4375,
+      "step": 2363
+    },
+    {
+      "epoch": 2.9006134969325155,
+      "grad_norm": 1.0131718797274043,
+      "learning_rate": 2.6776859504132232e-08,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -436.0,
+      "logps/rejected": -444.0,
+      "loss": 0.0019,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.291015625,
+      "rewards/margins": 13.375,
+      "rewards/rejected": -13.125,
+      "step": 2364
+    },
+    {
+      "epoch": 2.901840490797546,
+      "grad_norm": 2.721830783305495,
+      "learning_rate": 2.6446280991735537e-08,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0118,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.76171875,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.1875,
+      "step": 2365
+    },
+    {
+      "epoch": 2.903067484662577,
+      "grad_norm": 1.6130518393246316,
+      "learning_rate": 2.611570247933884e-08,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0082,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.71484375,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.75,
+      "step": 2366
+    },
+    {
+      "epoch": 2.9042944785276075,
+      "grad_norm": 1.312294080949433,
+      "learning_rate": 2.5785123966942146e-08,
+      "logits/chosen": -0.318359375,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -342.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1171875,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.875,
+      "step": 2367
+    },
+    {
+      "epoch": 2.905521472392638,
+      "grad_norm": 1.6646843333303585,
+      "learning_rate": 2.545454545454545e-08,
+      "logits/chosen": -0.392578125,
+      "logits/rejected": -0.58203125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -442.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.337890625,
+      "rewards/margins": 13.375,
+      "rewards/rejected": -13.6875,
+      "step": 2368
+    },
+    {
+      "epoch": 2.906748466257669,
+      "grad_norm": 1.491271563203498,
+      "learning_rate": 2.512396694214876e-08,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5703125,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.0,
+      "step": 2369
+    },
+    {
+      "epoch": 2.9079754601226995,
+      "grad_norm": 2.150753146030636,
+      "learning_rate": 2.4793388429752066e-08,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0041,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.62109375,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.75,
+      "step": 2370
+    },
+    {
+      "epoch": 2.90920245398773,
+      "grad_norm": 2.5014244365288967,
+      "learning_rate": 2.446280991735537e-08,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0092,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.28125,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -12.8125,
+      "step": 2371
+    },
+    {
+      "epoch": 2.910429447852761,
+      "grad_norm": 1.338260211121758,
+      "learning_rate": 2.4132231404958678e-08,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.6171875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.05322265625,
+      "rewards/margins": 13.625,
+      "rewards/rejected": -13.6875,
+      "step": 2372
+    },
+    {
+      "epoch": 2.9116564417177915,
+      "grad_norm": 3.6173168816798134,
+      "learning_rate": 2.3801652892561983e-08,
+      "logits/chosen": -0.10986328125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -436.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0061,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.47265625,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -12.625,
+      "step": 2373
+    },
+    {
+      "epoch": 2.912883435582822,
+      "grad_norm": 3.497704939335735,
+      "learning_rate": 2.3471074380165287e-08,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1279296875,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.375,
+      "step": 2374
+    },
+    {
+      "epoch": 2.914110429447853,
+      "grad_norm": 1.5123669301106812,
+      "learning_rate": 2.3140495867768595e-08,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -422.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.47265625,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.1875,
+      "step": 2375
+    },
+    {
+      "epoch": 2.9153374233128835,
+      "grad_norm": 2.6667528941285763,
+      "learning_rate": 2.28099173553719e-08,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.43359375,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.4375,
+      "step": 2376
+    },
+    {
+      "epoch": 2.916564417177914,
+      "grad_norm": 2.6827317858822646,
+      "learning_rate": 2.2479338842975207e-08,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -402.0,
+      "loss": 0.0063,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.88671875,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -12.3125,
+      "step": 2377
+    },
+    {
+      "epoch": 2.917791411042945,
+      "grad_norm": 1.1716330999719473,
+      "learning_rate": 2.214876033057851e-08,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.470703125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0084,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.93359375,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.75,
+      "step": 2378
+    },
+    {
+      "epoch": 2.9190184049079755,
+      "grad_norm": 2.067051974647425,
+      "learning_rate": 2.1818181818181816e-08,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -408.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.30078125,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.8125,
+      "step": 2379
+    },
+    {
+      "epoch": 2.920245398773006,
+      "grad_norm": 2.3002578798832145,
+      "learning_rate": 2.1487603305785124e-08,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.859375,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.125,
+      "step": 2380
+    },
+    {
+      "epoch": 2.921472392638037,
+      "grad_norm": 1.850438251488235,
+      "learning_rate": 2.1157024793388428e-08,
+      "logits/chosen": -0.3203125,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -410.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.7265625,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.125,
+      "step": 2381
+    },
+    {
+      "epoch": 2.9226993865030675,
+      "grad_norm": 1.5061458025258083,
+      "learning_rate": 2.0826446280991736e-08,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0091,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.57421875,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.875,
+      "step": 2382
+    },
+    {
+      "epoch": 2.923926380368098,
+      "grad_norm": 1.4436996047074342,
+      "learning_rate": 2.049586776859504e-08,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.412109375,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -12.75,
+      "step": 2383
+    },
+    {
+      "epoch": 2.925153374233129,
+      "grad_norm": 1.7674751230865213,
+      "learning_rate": 2.0165289256198345e-08,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0097,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.08349609375,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -12.625,
+      "step": 2384
+    },
+    {
+      "epoch": 2.9263803680981595,
+      "grad_norm": 3.7282848901058006,
+      "learning_rate": 1.9834710743801653e-08,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0079,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2890625,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.75,
+      "step": 2385
+    },
+    {
+      "epoch": 2.92760736196319,
+      "grad_norm": 1.0883425726395326,
+      "learning_rate": 1.9504132231404957e-08,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.462890625,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -11.6875,
+      "step": 2386
+    },
+    {
+      "epoch": 2.928834355828221,
+      "grad_norm": 1.4905192624885324,
+      "learning_rate": 1.9173553719008265e-08,
+      "logits/chosen": -0.341796875,
+      "logits/rejected": -0.53125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.81640625,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.4375,
+      "step": 2387
+    },
+    {
+      "epoch": 2.9300613496932515,
+      "grad_norm": 1.4261469909548503,
+      "learning_rate": 1.884297520661157e-08,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -420.0,
+      "logps/rejected": -416.0,
+      "loss": 0.003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.166015625,
+      "rewards/margins": 13.6875,
+      "rewards/rejected": -13.5,
+      "step": 2388
+    },
+    {
+      "epoch": 2.931288343558282,
+      "grad_norm": 3.5251275028334614,
+      "learning_rate": 1.8512396694214874e-08,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -418.0,
+      "loss": 0.0082,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.328125,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.0625,
+      "step": 2389
+    },
+    {
+      "epoch": 2.932515337423313,
+      "grad_norm": 2.0348926778673926,
+      "learning_rate": 1.818181818181818e-08,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.490234375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -404.0,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1611328125,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -12.625,
+      "step": 2390
+    },
+    {
+      "epoch": 2.9337423312883435,
+      "grad_norm": 1.4257611126488703,
+      "learning_rate": 1.7851239669421486e-08,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0084,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.56640625,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.4375,
+      "step": 2391
+    },
+    {
+      "epoch": 2.934969325153374,
+      "grad_norm": 4.841232499453999,
+      "learning_rate": 1.7520661157024794e-08,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -424.0,
+      "loss": 0.0062,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.51953125,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.25,
+      "step": 2392
+    },
+    {
+      "epoch": 2.936196319018405,
+      "grad_norm": 2.513915523771651,
+      "learning_rate": 1.7190082644628098e-08,
+      "logits/chosen": -0.2001953125,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -352.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0073,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.80078125,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.875,
+      "step": 2393
+    },
+    {
+      "epoch": 2.9374233128834355,
+      "grad_norm": 1.2072839440742664,
+      "learning_rate": 1.6859504132231403e-08,
+      "logits/chosen": -0.330078125,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -430.0,
+      "loss": 0.0079,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.39453125,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -13.1875,
+      "step": 2394
+    },
+    {
+      "epoch": 2.938650306748466,
+      "grad_norm": 2.4791544528621494,
+      "learning_rate": 1.652892561983471e-08,
+      "logits/chosen": -0.353515625,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.69140625,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -13.0625,
+      "step": 2395
+    },
+    {
+      "epoch": 2.939877300613497,
+      "grad_norm": 1.1991395052184106,
+      "learning_rate": 1.6198347107438015e-08,
+      "logits/chosen": -0.36328125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -436.0,
+      "loss": 0.0029,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -12.5625,
+      "step": 2396
+    },
+    {
+      "epoch": 2.9411042944785275,
+      "grad_norm": 1.809563258451705,
+      "learning_rate": 1.5867768595041323e-08,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.703125,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.0,
+      "step": 2397
+    },
+    {
+      "epoch": 2.942331288343558,
+      "grad_norm": 1.0570139089467685,
+      "learning_rate": 1.5537190082644627e-08,
+      "logits/chosen": -0.1396484375,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -408.0,
+      "loss": 0.0026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1015625,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -13.125,
+      "step": 2398
+    },
+    {
+      "epoch": 2.943558282208589,
+      "grad_norm": 1.7855014046378934,
+      "learning_rate": 1.5206611570247932e-08,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.75,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -13.0625,
+      "step": 2399
+    },
+    {
+      "epoch": 2.9447852760736195,
+      "grad_norm": 2.5585400504509814,
+      "learning_rate": 1.487603305785124e-08,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -430.0,
+      "logps/rejected": -434.0,
+      "loss": 0.0043,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.54296875,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -13.6875,
+      "step": 2400
+    },
+    {
+      "epoch": 2.94601226993865,
+      "grad_norm": 4.055719353360522,
+      "learning_rate": 1.4545454545454544e-08,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.490234375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -388.0,
+      "loss": 0.0109,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -0.302734375,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.75,
+      "step": 2401
+    },
+    {
+      "epoch": 2.947239263803681,
+      "grad_norm": 3.553567636626623,
+      "learning_rate": 1.4214876033057852e-08,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.201171875,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -13.0625,
+      "step": 2402
+    },
+    {
+      "epoch": 2.9484662576687115,
+      "grad_norm": 3.256536448489545,
+      "learning_rate": 1.3884297520661156e-08,
+      "logits/chosen": -0.326171875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.78515625,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.0,
+      "step": 2403
+    },
+    {
+      "epoch": 2.949693251533742,
+      "grad_norm": 4.132877653562335,
+      "learning_rate": 1.355371900826446e-08,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -442.0,
+      "logps/rejected": -430.0,
+      "loss": 0.006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.19921875,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -12.6875,
+      "step": 2404
+    },
+    {
+      "epoch": 2.950920245398773,
+      "grad_norm": 3.8004168993428067,
+      "learning_rate": 1.3223140495867768e-08,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0055,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.05712890625,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -11.875,
+      "step": 2405
+    },
+    {
+      "epoch": 2.9521472392638035,
+      "grad_norm": 2.603038109739486,
+      "learning_rate": 1.2892561983471073e-08,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -396.0,
+      "loss": 0.0059,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1015625,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.8125,
+      "step": 2406
+    },
+    {
+      "epoch": 2.9533742331288346,
+      "grad_norm": 2.150911418844657,
+      "learning_rate": 1.256198347107438e-08,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.337890625,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.25,
+      "step": 2407
+    },
+    {
+      "epoch": 2.954601226993865,
+      "grad_norm": 3.3512675351177403,
+      "learning_rate": 1.2231404958677685e-08,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -432.0,
+      "loss": 0.0073,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.330078125,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -12.875,
+      "step": 2408
+    },
+    {
+      "epoch": 2.955828220858896,
+      "grad_norm": 1.265596062357397,
+      "learning_rate": 1.1900826446280991e-08,
+      "logits/chosen": -0.484375,
+      "logits/rejected": -0.640625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.06298828125,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.125,
+      "step": 2409
+    },
+    {
+      "epoch": 2.957055214723926,
+      "grad_norm": 1.5869524875524,
+      "learning_rate": 1.1570247933884297e-08,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -400.0,
+      "logps/rejected": -398.0,
+      "loss": 0.0044,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.61328125,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.3125,
+      "step": 2410
+    },
+    {
+      "epoch": 2.9582822085889573,
+      "grad_norm": 1.341909650879635,
+      "learning_rate": 1.1239669421487603e-08,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0039,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -12.625,
+      "step": 2411
+    },
+    {
+      "epoch": 2.9595092024539875,
+      "grad_norm": 2.109735694441826,
+      "learning_rate": 1.0909090909090908e-08,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -438.0,
+      "loss": 0.0048,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.60546875,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.875,
+      "step": 2412
+    },
+    {
+      "epoch": 2.9607361963190186,
+      "grad_norm": 1.113936073329194,
+      "learning_rate": 1.0578512396694214e-08,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -462.0,
+      "logps/rejected": -446.0,
+      "loss": 0.0024,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.09716796875,
+      "rewards/margins": 14.1875,
+      "rewards/rejected": -14.125,
+      "step": 2413
+    },
+    {
+      "epoch": 2.961963190184049,
+      "grad_norm": 2.968058393263657,
+      "learning_rate": 1.024793388429752e-08,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -320.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0063,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.53125,
+      "rewards/margins": 10.3125,
+      "rewards/rejected": -11.875,
+      "step": 2414
+    },
+    {
+      "epoch": 2.96319018404908,
+      "grad_norm": 4.877190166655265,
+      "learning_rate": 9.917355371900826e-09,
+      "logits/chosen": -0.44140625,
+      "logits/rejected": -0.65234375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0066,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.361328125,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.625,
+      "step": 2415
+    },
+    {
+      "epoch": 2.96441717791411,
+      "grad_norm": 1.9939936656295827,
+      "learning_rate": 9.586776859504132e-09,
+      "logits/chosen": -0.361328125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -452.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.267578125,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -13.5625,
+      "step": 2416
+    },
+    {
+      "epoch": 2.9656441717791413,
+      "grad_norm": 4.349077654491785,
+      "learning_rate": 9.256198347107437e-09,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -420.0,
+      "loss": 0.0077,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2578125,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -13.1875,
+      "step": 2417
+    },
+    {
+      "epoch": 2.9668711656441715,
+      "grad_norm": 1.3690177008162818,
+      "learning_rate": 8.925619834710743e-09,
+      "logits/chosen": -0.341796875,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -418.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.162109375,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -12.5625,
+      "step": 2418
+    },
+    {
+      "epoch": 2.9680981595092026,
+      "grad_norm": 3.5886541807689616,
+      "learning_rate": 8.595041322314049e-09,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -410.0,
+      "loss": 0.0065,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.79296875,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -13.0,
+      "step": 2419
+    },
+    {
+      "epoch": 2.969325153374233,
+      "grad_norm": 2.0968189514243383,
+      "learning_rate": 8.264462809917355e-09,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0047,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.330078125,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.75,
+      "step": 2420
+    },
+    {
+      "epoch": 2.970552147239264,
+      "grad_norm": 4.0764634657112415,
+      "learning_rate": 7.933884297520661e-09,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -406.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.322265625,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.0625,
+      "step": 2421
+    },
+    {
+      "epoch": 2.9717791411042946,
+      "grad_norm": 2.0254198433359254,
+      "learning_rate": 7.603305785123966e-09,
+      "logits/chosen": -0.40625,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -426.0,
+      "loss": 0.004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.796875,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.0625,
+      "step": 2422
+    },
+    {
+      "epoch": 2.9730061349693253,
+      "grad_norm": 2.6750387904939017,
+      "learning_rate": 7.272727272727272e-09,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -378.0,
+      "loss": 0.0056,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.55859375,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -12.875,
+      "step": 2423
+    },
+    {
+      "epoch": 2.974233128834356,
+      "grad_norm": 1.204893450531445,
+      "learning_rate": 6.942148760330578e-09,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0031,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.265625,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -13.0,
+      "step": 2424
+    },
+    {
+      "epoch": 2.9754601226993866,
+      "grad_norm": 2.767561503985594,
+      "learning_rate": 6.611570247933884e-09,
+      "logits/chosen": -0.263671875,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -440.0,
+      "loss": 0.0049,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.25390625,
+      "rewards/margins": 13.5625,
+      "rewards/rejected": -13.3125,
+      "step": 2425
+    },
+    {
+      "epoch": 2.9766871165644173,
+      "grad_norm": 1.3223051040251024,
+      "learning_rate": 6.28099173553719e-09,
+      "logits/chosen": -0.431640625,
+      "logits/rejected": -0.625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -400.0,
+      "loss": 0.0028,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.68359375,
+      "rewards/margins": 13.4375,
+      "rewards/rejected": -14.0625,
+      "step": 2426
+    },
+    {
+      "epoch": 2.977914110429448,
+      "grad_norm": 1.2794031404252872,
+      "learning_rate": 5.950413223140496e-09,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -384.0,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.921875,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.1875,
+      "step": 2427
+    },
+    {
+      "epoch": 2.9791411042944786,
+      "grad_norm": 1.0131054877574335,
+      "learning_rate": 5.619834710743802e-09,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.435546875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -386.0,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1416015625,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -12.3125,
+      "step": 2428
+    },
+    {
+      "epoch": 2.9803680981595093,
+      "grad_norm": 1.866346659152866,
+      "learning_rate": 5.289256198347107e-09,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -412.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.76171875,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.375,
+      "step": 2429
+    },
+    {
+      "epoch": 2.98159509202454,
+      "grad_norm": 1.7846273867699272,
+      "learning_rate": 4.958677685950413e-09,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0034,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.384765625,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.0,
+      "step": 2430
+    },
+    {
+      "epoch": 2.9828220858895707,
+      "grad_norm": 1.0077622063356493,
+      "learning_rate": 4.6280991735537185e-09,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -428.0,
+      "loss": 0.0019,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1552734375,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -13.125,
+      "step": 2431
+    },
+    {
+      "epoch": 2.9840490797546013,
+      "grad_norm": 1.6234181675136565,
+      "learning_rate": 4.2975206611570246e-09,
+      "logits/chosen": -0.3359375,
+      "logits/rejected": -0.4921875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -438.0,
+      "loss": 0.009,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.90625,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -13.625,
+      "step": 2432
+    },
+    {
+      "epoch": 2.985276073619632,
+      "grad_norm": 1.2855678287996661,
+      "learning_rate": 3.966942148760331e-09,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.53515625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -396.0,
+      "loss": 0.008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.064453125,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.9375,
+      "step": 2433
+    },
+    {
+      "epoch": 2.9865030674846627,
+      "grad_norm": 1.062117879895051,
+      "learning_rate": 3.636363636363636e-09,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.61328125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -394.0,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.71875,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -12.0,
+      "step": 2434
+    },
+    {
+      "epoch": 2.9877300613496933,
+      "grad_norm": 2.4006989852791363,
+      "learning_rate": 3.305785123966942e-09,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0046,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0078125,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -12.0625,
+      "step": 2435
+    },
+    {
+      "epoch": 2.988957055214724,
+      "grad_norm": 3.825713501073963,
+      "learning_rate": 2.975206611570248e-09,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -486.0,
+      "logps/rejected": -422.0,
+      "loss": 0.0064,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.58984375,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -12.625,
+      "step": 2436
+    },
+    {
+      "epoch": 2.9901840490797547,
+      "grad_norm": 1.6694083195462188,
+      "learning_rate": 2.6446280991735535e-09,
+      "logits/chosen": -0.341796875,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -444.0,
+      "loss": 0.0042,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2158203125,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -12.9375,
+      "step": 2437
+    },
+    {
+      "epoch": 2.9914110429447853,
+      "grad_norm": 0.9933305662734105,
+      "learning_rate": 2.3140495867768592e-09,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -382.0,
+      "loss": 0.0021,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.5078125,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -12.75,
+      "step": 2438
+    },
+    {
+      "epoch": 2.992638036809816,
+      "grad_norm": 1.803380952459384,
+      "learning_rate": 1.9834710743801653e-09,
+      "logits/chosen": -0.302734375,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.051025390625,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.125,
+      "step": 2439
+    },
+    {
+      "epoch": 2.9938650306748467,
+      "grad_norm": 0.943324091134692,
+      "learning_rate": 1.652892561983471e-09,
+      "logits/chosen": -0.4765625,
+      "logits/rejected": -0.55859375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -460.0,
+      "loss": 0.0076,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": 0.07080078125,
+      "rewards/margins": 14.0,
+      "rewards/rejected": -13.9375,
+      "step": 2440
+    },
+    {
+      "epoch": 2.9950920245398773,
+      "grad_norm": 1.795962559751925,
+      "learning_rate": 1.3223140495867768e-09,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0104,
+      "rewards/accuracies": 0.9921875,
+      "rewards/chosen": -1.4765625,
+      "rewards/margins": 10.625,
+      "rewards/rejected": -12.125,
+      "step": 2441
+    },
+    {
+      "epoch": 2.996319018404908,
+      "grad_norm": 1.1010491562461566,
+      "learning_rate": 9.917355371900827e-10,
+      "logits/chosen": -0.2421875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -414.0,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1376953125,
+      "rewards/margins": 13.0625,
+      "rewards/rejected": -12.875,
+      "step": 2442
+    },
+    {
+      "epoch": 2.9975460122699387,
+      "grad_norm": 1.7985189665374313,
+      "learning_rate": 6.611570247933884e-10,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0037,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.5,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -12.875,
+      "step": 2443
+    },
+    {
+      "epoch": 2.9987730061349693,
+      "grad_norm": 3.1261485435103427,
+      "learning_rate": 3.305785123966942e-10,
+      "logits/chosen": -0.4140625,
+      "logits/rejected": -0.6015625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0051,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.44921875,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.0625,
+      "step": 2444
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.7705045301693025,
+      "learning_rate": 0.0,
+      "logits/chosen": -0.400390625,
+      "logits/rejected": -0.5390625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -426.0,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.0216064453125,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -12.875,
+      "step": 2445
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2445,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-2445/training_args.bin b/checkpoint-2445/training_args.bin
new file mode 100644
index 0000000..70c7ec2
--- /dev/null
+++ b/checkpoint-2445/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:528936595b6d96d00e537c7ae1257bb1454c1a31a8f47041035d37b1fc6d35e1
+size 9528
diff --git a/checkpoint-2445/zero_to_fp32.py b/checkpoint-2445/zero_to_fp32.py
new file mode 100644
index 0000000..24cc342
--- /dev/null
+++ b/checkpoint-2445/zero_to_fp32.py
@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoint-815/config.json b/checkpoint-815/config.json
new file mode 100644
index 0000000..ded5a8b
--- /dev/null
+++ b/checkpoint-815/config.json
@@ -0,0 +1,35 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128009,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 8.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.50.0",
+  "use_cache": false,
+  "vocab_size": 128256
+}
diff --git a/checkpoint-815/generation_config.json b/checkpoint-815/generation_config.json
new file mode 100644
index 0000000..0acdb5a
--- /dev/null
+++ b/checkpoint-815/generation_config.json
@@ -0,0 +1,12 @@
+{
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.50.0"
+}
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..b5708cd
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f2f7bdeb356122f7aace8c3296690bedc8e4727fd59cb3780b9c3725c5d27f0
+size 1514015111
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..5563c18
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ae3652b3e9fece684b917e0e78997a28d1f4bc0c07d2a979dd449efa5a65bf5
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..d065ce5
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:144457185371d5dee6cbed51f6210bf146644934bd4feb0a6bc3d5707e55c52a
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..eaa566d
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d1fc011894154f5d0d60e030056b46861989f01e64e6e622c6ff1229ff62d38
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..949f5ef
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94d82f2adcd26dca6257a94e351b5c5f0c6f3c6e17a6ae283d841450bdbc7b10
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..26337c3
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:940dbd16742c5009426c4b1d581468f8fbdb545b3b39a7315b22a1652002b514
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..2993f4c
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df973ce6e5adcf353299b5aebdde777e8e56bd9163d961fcff7424da54df52c4
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..d161cc7
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa659ff40d3e33a83012d7e94b087ae633624c4c73b9db2d1e6c3e6f311285d3
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..bd8a0a2
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa93f73ea80b8278d1108bc23fef3c44327329924996058b721896f28debfcb8
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..cf1ef0f
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a97eaddf75945b1c69d758d42eeb2d6ab9c62e4ed15d84d37bcfda6ae140f2ef
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..0e7d882
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6999a55c6c8b045223e5d2b12f3b73d8fad2ace1c9d2550ca5dd9282b6fe9511
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..9c0ffd0
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b313bd450529f7746356c181adbd7b087d86646d406613933d36403a1b2afb9
+size 1514015111
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..e12c52a
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:931bc788618b58acbacb208f6534f9b08d3223dfa4607990d367ed0f5c1cd3f8
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..cad98ee
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6dd988b19007e5a6b46eb96354809de6352ce363927f3543ef0fa0a79ea5084d
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..baa2c64
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f95a042af40f785247882008985248e734d67023a4a598ae5f8386c75497807
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..3288a6d
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf68d2f38f5b4a8ed8b486cc6c00b7ce74b0b4f8cfa7ec39e3a930c6d3c92acb
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..ce1dc8d
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4065cfa5848223ffc3ca04624f717151f11f04d43bbdbe20e0df19a979b03914
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..79c710b
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3958d6b67d15baa312b11e6556655c245fcd4b4d99fb61ff665a567457f55141
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..f80a4cd
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eeb0a5ea1c70d77fd9fc8bde9bdec3dface9098193bc4ee8f56377e29a108e2b
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..9f1771b
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f236558e077036718f31b0b35b6192c1d54e669ba2d9fb8d28034cbfec9832c6
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..e5f3e39
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45ad9710171ab15336eda1b638482c9998550baf9fdb4efb8851539162f888d3
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..79e75ca
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94c84f1a2b261934de7a1af24051c30e1682823f5381b0a4482a3d1ef94fda8e
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..1807ec3
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f367a501c51c9684db72c272465833f72e23ac79e766bac757f5233e4cb1e53
+size 1514015111
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..d73fa5e
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e3ea061324d02b4c3b2c51d53fab4ef0068794c6a04397f579017511bfd74a6
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..029735c
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac54c7e5d0734400649d0e35db653ca65be9d5eaf081b71cf5527e536f0e3609
+size 1514016442
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..3b527bc
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e69eec335096831c5ccfd313f3cd183e49746126613f35b697c8473f4c74b94f
+size 1514015111
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..6855316
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0466e4e43eb81b81120ba316ea4fa63b2415e42870429b46d8a5f6da185f8c76
+size 1514015111
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..aa8870b
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c508b703ef594684ea0f65460c61544fa404e65f73b2a41b38b9385136f7de12
+size 1514015111
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..57dd569
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c7241f14f434f922428cd3a951c157907e35c0e5e8867f6990e2320b81cdc20
+size 1514015111
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..11847b7
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1318c4542cd7ea51e207b7c74afb105ebb3e29744fbbb300f95a168a84fb7ea9
+size 1514015111
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..a532388
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27161fec1e8389687a78630c37bc20776c8cc0e6f5f40a5ce4846a3d09464301
+size 1514015111
diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt
new file mode 100644
index 0000000..178a137
--- /dev/null
+++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a62d552afba1ded3f321836180156bfeecfd3934378808d7389fa673ffcfbde
+size 1514015111
diff --git a/checkpoint-815/global_step815/zero_pp_rank_0_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_0_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..8948f42
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_0_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40ba6c0d314a71a8a038eeffa9046553c73f5e76e4b9d8e856629862855dfcfb
+size 150245
diff --git a/checkpoint-815/global_step815/zero_pp_rank_10_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_10_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..2212482
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_10_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:024eef49228ce027257c81680e51700d64008ec55789eb59dcde70a36817558a
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_11_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_11_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..40bd1bd
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_11_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c00f3b3e366b7cfacd447b4434da21ef301448603fb55d4335db39299b66575
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_12_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_12_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..9915c4d
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_12_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6cf1da6bf30a5b8f93f325e72d167effb11dfdf29766364fef6b9d4506dfb81e
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_13_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_13_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..f03df1f
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_13_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f94229da5713f438ab5c4189e38eadbc168fc643fd01a5b6eb930bd2f722efd
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_14_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_14_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..058ff88
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_14_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b5b97915295b3f6c7eca69f6541ba5833b8919fc0cb698d51f36e8f4b9fdf1d5
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_15_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_15_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..b950999
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_15_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b99f4e99e6ffd910f61d39cd0667ad6f7cf15a6179452f54cd0ab9eb3be0f7c9
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_16_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_16_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..8974561
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_16_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61fef944ea931de77863bf7ca7b70cda46266397ec948e9282e6930a74bf2ed0
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_17_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_17_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..05f05a1
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_17_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c65e902223780bcaa882920de3575f35c5deb86a74decf182ff60de992138c6
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_18_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_18_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..238e639
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_18_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a133f1cb598eed313d7415e6de2861d456036a9823785dafc9b52a26fa533f4c
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_19_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_19_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..2c4dc0d
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_19_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55591d7a7ad8cacac5cf43918a7ad42f693db1b850f57a260cf2778aee083cef
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_1_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_1_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..17bb96a
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_1_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5947cccfabfeb8fc303b07ca84c4d4146bb568b41e8dee350a89b8dbb72981c
+size 150245
diff --git a/checkpoint-815/global_step815/zero_pp_rank_20_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_20_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..1604c21
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_20_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7c17e090edef85f94359ec12ac733fd9394d7a36675b4cf63dda67fe6da8f24
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_21_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_21_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..98d5fce
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_21_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d35bbd41a45c012aa1db3445f9b72b3dd567ebc41b00dce74bd1e4d82e2ebfd
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_22_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_22_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..91cd1fe
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_22_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e38f525185b44ebf31ebd0ac8f0d8a32e0fb510d3f674311c087e3508b40c993
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_23_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_23_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..168e223
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_23_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f7cce2fd60981a15a0f3f33cc90ce2537fcaab35671f44f3c49fcea6fca1891
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_24_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_24_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..ab22b84
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_24_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f8d1850a9fd50d9ad858ec306dadb3409f83a743868167753d2ceb1ee0b8782
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_25_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_25_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..3d155ea
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_25_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55bc11c638b6fadef2d46875c4a0e4a788f2b78ae95eacdff3a036f1cbfc8b9e
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_26_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_26_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..32b8482
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_26_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff219377c7fb54bb953c6c570cf957b0e25d7bb9d2326d87143bc2db6e131b02
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_27_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_27_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..ad8f112
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_27_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18ed23baa1f8cd3b5be593df62b46bd925963eae398fb856c46af8bca0bdc706
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_28_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_28_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..572add1
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_28_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e4fe2622f3c42b1d9b6319829c12a40ba41eed4b5165643ba1e541bbd9d6773
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_29_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_29_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..8f66a39
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_29_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3aaab1637c614043c25c7321822568b135d719fb8ee06fdcd666152646a036f8
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_2_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_2_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..bce022e
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_2_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:955cdff9e07f053cbc74b4a713e4d8def17fa2e19b7e9802bad227f51d4b61af
+size 150245
diff --git a/checkpoint-815/global_step815/zero_pp_rank_30_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_30_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..d78e9d3
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_30_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc85012ee362480a408246d70a9f1fce990efdf194b19f0a18f7f3aac1eb1ab3
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_31_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_31_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..025216a
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_31_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2fbdfe073710344bded771325bbd465f5b8a32a32bdf561a674751b41cf6e111
+size 150540
diff --git a/checkpoint-815/global_step815/zero_pp_rank_3_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_3_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..45a8ff1
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_3_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1bbe876d4201139b9ab1ecc5cc62f74e0eb73ef120e1cff4c5f84dca3059dad7
+size 150245
diff --git a/checkpoint-815/global_step815/zero_pp_rank_4_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_4_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..51ec98a
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_4_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bfae4502edaad044ae8c1a527900b97fd20fea07eaa512af7273532d23b47618
+size 150245
diff --git a/checkpoint-815/global_step815/zero_pp_rank_5_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_5_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..32e7ce4
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_5_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73170415f2c2bef296c8730bf78811e31654ac55c87c2d71d77bd88d30b5a270
+size 150245
diff --git a/checkpoint-815/global_step815/zero_pp_rank_6_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_6_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..82d10da
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_6_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:859c2393bcde923cc30a036b27c836d6eac277f575f022b4cbb0b541237091dd
+size 150245
diff --git a/checkpoint-815/global_step815/zero_pp_rank_7_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_7_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..00c0875
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_7_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4494a1f9263497b860dafc470670372cabcf493bd54adb561d84e9c57be3dac7
+size 150245
diff --git a/checkpoint-815/global_step815/zero_pp_rank_8_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_8_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..aee6ef3
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_8_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:edb26fdb729d3f1fea2b31878462cdc9057312c888b9c8ef8bae28e1002dfdf4
+size 150245
diff --git a/checkpoint-815/global_step815/zero_pp_rank_9_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_9_mp_rank_00_model_states.pt
new file mode 100644
index 0000000..077564d
--- /dev/null
+++ b/checkpoint-815/global_step815/zero_pp_rank_9_mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22843c3e9c8f1d5116a8f88a79ac3c5368fdf132311c803b3b1b2d75ea9cb1b6
+size 150245
diff --git a/checkpoint-815/latest b/checkpoint-815/latest
new file mode 100644
index 0000000..c057bc6
--- /dev/null
+++ b/checkpoint-815/latest
@@ -0,0 +1 @@
+global_step815
\ No newline at end of file
diff --git a/checkpoint-815/model-00001-of-00004.safetensors b/checkpoint-815/model-00001-of-00004.safetensors
new file mode 100644
index 0000000..2313115
--- /dev/null
+++ b/checkpoint-815/model-00001-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd0460b7a7e5f80b033a47d66628276be359c8c130b8e2556b7ce7b034e73d8f
+size 4976698672
diff --git a/checkpoint-815/model-00002-of-00004.safetensors b/checkpoint-815/model-00002-of-00004.safetensors
new file mode 100644
index 0000000..14885e9
--- /dev/null
+++ b/checkpoint-815/model-00002-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3b878c898c4556c9c80a89237f335fe01f816bb0811471532ec74e95671ea581
+size 4999802720
diff --git a/checkpoint-815/model-00003-of-00004.safetensors b/checkpoint-815/model-00003-of-00004.safetensors
new file mode 100644
index 0000000..a7f8cca
--- /dev/null
+++ b/checkpoint-815/model-00003-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3ef0b22e38f27de4b856a8ce555608a5c3b623be93b9d876e32ace6f499c202
+size 4915916176
diff --git a/checkpoint-815/model-00004-of-00004.safetensors b/checkpoint-815/model-00004-of-00004.safetensors
new file mode 100644
index 0000000..f4495d0
--- /dev/null
+++ b/checkpoint-815/model-00004-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d39f7385fcb7dcbc2e69697038c9b983c028d240ed0da2e27c989556b194696
+size 1168138808
diff --git a/checkpoint-815/model.safetensors.index.json b/checkpoint-815/model.safetensors.index.json
new file mode 100644
index 0000000..0fd8120
--- /dev/null
+++ b/checkpoint-815/model.safetensors.index.json
@@ -0,0 +1,298 @@
+{
+  "metadata": {
+    "total_size": 16060522496
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.norm.weight": "model-00004-of-00004.safetensors"
+  }
+}
diff --git a/checkpoint-815/rng_state_0.pth b/checkpoint-815/rng_state_0.pth
new file mode 100644
index 0000000..b647361
--- /dev/null
+++ b/checkpoint-815/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:575119a228f98110923ffa2dedcb50e3317251b26054355d015e0b2240d566f2
+size 15984
diff --git a/checkpoint-815/rng_state_1.pth b/checkpoint-815/rng_state_1.pth
new file mode 100644
index 0000000..8506e00
--- /dev/null
+++ b/checkpoint-815/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0728b56dab7abb5ef8a0d4bae3519c5767c97467bdd886d26bf19cc8599d0312
+size 15984
diff --git a/checkpoint-815/rng_state_10.pth b/checkpoint-815/rng_state_10.pth
new file mode 100644
index 0000000..4a20fa5
--- /dev/null
+++ b/checkpoint-815/rng_state_10.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:532d1ab071177c2e9d0cc3435d6314c1777a0356f9ce1a5d8f10fbe5f8fd8926
+size 15997
diff --git a/checkpoint-815/rng_state_11.pth b/checkpoint-815/rng_state_11.pth
new file mode 100644
index 0000000..a0cf51d
--- /dev/null
+++ b/checkpoint-815/rng_state_11.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a089513416d3ed447265187782ec1194ac2303155b941284463976c62bf3d9b
+size 15997
diff --git a/checkpoint-815/rng_state_12.pth b/checkpoint-815/rng_state_12.pth
new file mode 100644
index 0000000..97aa200
--- /dev/null
+++ b/checkpoint-815/rng_state_12.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a87db6b16c9956496c45bf5c81750dc2349e157e60e544e5d256bb957be1243
+size 15997
diff --git a/checkpoint-815/rng_state_13.pth b/checkpoint-815/rng_state_13.pth
new file mode 100644
index 0000000..c9f4510
--- /dev/null
+++ b/checkpoint-815/rng_state_13.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36d5fe5d1b92643a3c3dd35cf2f0cbfa4a92599fe3ccbd8941395c1550f088aa
+size 15997
diff --git a/checkpoint-815/rng_state_14.pth b/checkpoint-815/rng_state_14.pth
new file mode 100644
index 0000000..3c9daf6
--- /dev/null
+++ b/checkpoint-815/rng_state_14.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4a08278e808af796415aac256bf0c1e8619db1f0f025ff15775cd3b8d02bd59
+size 15997
diff --git a/checkpoint-815/rng_state_15.pth b/checkpoint-815/rng_state_15.pth
new file mode 100644
index 0000000..1697bc6
--- /dev/null
+++ b/checkpoint-815/rng_state_15.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88600b3520001a5a7e82c0e184d278b46492fda5424ea22ea243af15b40b2e82
+size 15997
diff --git a/checkpoint-815/rng_state_16.pth b/checkpoint-815/rng_state_16.pth
new file mode 100644
index 0000000..480705c
--- /dev/null
+++ b/checkpoint-815/rng_state_16.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9a5825e92eb7518856e88ce648529dc0597c32a5f52764f0361a119217a4dfb
+size 15997
diff --git a/checkpoint-815/rng_state_17.pth b/checkpoint-815/rng_state_17.pth
new file mode 100644
index 0000000..d425906
--- /dev/null
+++ b/checkpoint-815/rng_state_17.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b30bd2cfa327fb55a93d390d852db109e0a2e1f102ecabf520870cbc79ca95ad
+size 15997
diff --git a/checkpoint-815/rng_state_18.pth b/checkpoint-815/rng_state_18.pth
new file mode 100644
index 0000000..a8bc0d6
--- /dev/null
+++ b/checkpoint-815/rng_state_18.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77c95f2456fa4473e7231ab94b79ecbc3c126d31d3a94c61f86ee6125b547456
+size 15997
diff --git a/checkpoint-815/rng_state_19.pth b/checkpoint-815/rng_state_19.pth
new file mode 100644
index 0000000..40fe188
--- /dev/null
+++ b/checkpoint-815/rng_state_19.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:301f380300f7ae9a581c304ac407794ecff9d66e6c9c55ff6cffcfc8d94690a4
+size 15997
diff --git a/checkpoint-815/rng_state_2.pth b/checkpoint-815/rng_state_2.pth
new file mode 100644
index 0000000..ea499e2
--- /dev/null
+++ b/checkpoint-815/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4e481d4ef1546694da7337f6bb6c658b866dcb79b85deeb477da0d27ebe851e
+size 15984
diff --git a/checkpoint-815/rng_state_20.pth b/checkpoint-815/rng_state_20.pth
new file mode 100644
index 0000000..4b90c8b
--- /dev/null
+++ b/checkpoint-815/rng_state_20.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b2e90afb8f9481396fa1f82913f7fb908e83e63bf2a38a6f6d2b869ca1d06db
+size 15997
diff --git a/checkpoint-815/rng_state_21.pth b/checkpoint-815/rng_state_21.pth
new file mode 100644
index 0000000..fb87d85
--- /dev/null
+++ b/checkpoint-815/rng_state_21.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35fafcbe672a0f62668f219b7ae7b88ddc86330086a2eca82505408964d61498
+size 15997
diff --git a/checkpoint-815/rng_state_22.pth b/checkpoint-815/rng_state_22.pth
new file mode 100644
index 0000000..ac9beaf
--- /dev/null
+++ b/checkpoint-815/rng_state_22.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b8f851ca87ff60251262afa155c3ab50fc55615fb72c857120fd7fa3fe3177c
+size 15997
diff --git a/checkpoint-815/rng_state_23.pth b/checkpoint-815/rng_state_23.pth
new file mode 100644
index 0000000..59e574d
--- /dev/null
+++ b/checkpoint-815/rng_state_23.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27d8cd100c48bfcfbf5c8a5164b02d5879f69b5a6276dae705a9c60de9a079c1
+size 15997
diff --git a/checkpoint-815/rng_state_24.pth b/checkpoint-815/rng_state_24.pth
new file mode 100644
index 0000000..5064859
--- /dev/null
+++ b/checkpoint-815/rng_state_24.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78d1a58c82846308cc3df216d63b5388e6910766e8fbc6b4e8b28255bfff9f0b
+size 15997
diff --git a/checkpoint-815/rng_state_25.pth b/checkpoint-815/rng_state_25.pth
new file mode 100644
index 0000000..c0ea34f
--- /dev/null
+++ b/checkpoint-815/rng_state_25.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bdee831f657ca735b9fcf1c4034be6dcc25cc224594a3b2c7488ea6ec175be96
+size 15997
diff --git a/checkpoint-815/rng_state_26.pth b/checkpoint-815/rng_state_26.pth
new file mode 100644
index 0000000..f7f04e9
--- /dev/null
+++ b/checkpoint-815/rng_state_26.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e922469f65021a6179e6ecc7c2713db0a41dead9dc08d91e98ed34e65afd14e
+size 15997
diff --git a/checkpoint-815/rng_state_27.pth b/checkpoint-815/rng_state_27.pth
new file mode 100644
index 0000000..f354e1d
--- /dev/null
+++ b/checkpoint-815/rng_state_27.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11867e8beaaaacbefb60c912ca48d2026f0222b2247dd97d78a6ddd227e75bcf
+size 15997
diff --git a/checkpoint-815/rng_state_28.pth b/checkpoint-815/rng_state_28.pth
new file mode 100644
index 0000000..12c799a
--- /dev/null
+++ b/checkpoint-815/rng_state_28.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7664653b5278a30fcc7bd77ee73888988040ecbb6a9cb71fd8a5bb5ccd2c9a31
+size 15997
diff --git a/checkpoint-815/rng_state_29.pth b/checkpoint-815/rng_state_29.pth
new file mode 100644
index 0000000..6feba08
--- /dev/null
+++ b/checkpoint-815/rng_state_29.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5007396f0c3c95678e1e62a10e31552d6bfe2303e6e78ab6882752474b2efdc7
+size 15997
diff --git a/checkpoint-815/rng_state_3.pth b/checkpoint-815/rng_state_3.pth
new file mode 100644
index 0000000..aeb38f9
--- /dev/null
+++ b/checkpoint-815/rng_state_3.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:353c60be37ea56fc992fca446598ceca5d1fd002aa3bd6dbb9ad740e6f47ebb3
+size 15984
diff --git a/checkpoint-815/rng_state_30.pth b/checkpoint-815/rng_state_30.pth
new file mode 100644
index 0000000..52254ce
--- /dev/null
+++ b/checkpoint-815/rng_state_30.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f01c58f06cc2a38c9093fadd85def1f922a5e8910e17a0d885f0df5570b13a98
+size 15997
diff --git a/checkpoint-815/rng_state_31.pth b/checkpoint-815/rng_state_31.pth
new file mode 100644
index 0000000..f6010f3
--- /dev/null
+++ b/checkpoint-815/rng_state_31.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35e075e15ab066bd7c9405a66e387f2ebf6b302fbd19ba7905da3f8cc8f71009
+size 15997
diff --git a/checkpoint-815/rng_state_4.pth b/checkpoint-815/rng_state_4.pth
new file mode 100644
index 0000000..9d5856c
--- /dev/null
+++ b/checkpoint-815/rng_state_4.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9107fe964ba7205e354084b85210e5a5ea1c98cfd4d38adb9cd3926945dcae4
+size 15984
diff --git a/checkpoint-815/rng_state_5.pth b/checkpoint-815/rng_state_5.pth
new file mode 100644
index 0000000..b824ee2
--- /dev/null
+++ b/checkpoint-815/rng_state_5.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69d1bb1abee38b92e53f3f23549b642ce0f1edcdccf7b6129847ac61636e96d5
+size 15984
diff --git a/checkpoint-815/rng_state_6.pth b/checkpoint-815/rng_state_6.pth
new file mode 100644
index 0000000..a9fd036
--- /dev/null
+++ b/checkpoint-815/rng_state_6.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afd5516048e20f36959601574e29e40106085a7d3cdc7bf425ce5e84633490e6
+size 15984
diff --git a/checkpoint-815/rng_state_7.pth b/checkpoint-815/rng_state_7.pth
new file mode 100644
index 0000000..4e80125
--- /dev/null
+++ b/checkpoint-815/rng_state_7.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e2c46927fc06939b4c976a01e4b95dec1f8b98ceaea86d31a5d756fc30ff006
+size 15984
diff --git a/checkpoint-815/rng_state_8.pth b/checkpoint-815/rng_state_8.pth
new file mode 100644
index 0000000..b95c8ee
--- /dev/null
+++ b/checkpoint-815/rng_state_8.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87544a1ecb93bba51fc38c978d8a027be1635c93e80f7f0679c1154bc76e16d3
+size 15984
diff --git a/checkpoint-815/rng_state_9.pth b/checkpoint-815/rng_state_9.pth
new file mode 100644
index 0000000..8ff6eb6
--- /dev/null
+++ b/checkpoint-815/rng_state_9.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f51e1e81143bcfd2cfd36874b3b7606bb4a14aac945154b167b31a816d9f450
+size 15984
diff --git a/checkpoint-815/scheduler.pt b/checkpoint-815/scheduler.pt
new file mode 100644
index 0000000..cdbf7b3
--- /dev/null
+++ b/checkpoint-815/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17b4f8ffededff616b62d7ac6b7cffb068dd0871affa07c314815c7e89ea5e23
+size 1064
diff --git a/checkpoint-815/special_tokens_map.json b/checkpoint-815/special_tokens_map.json
new file mode 100644
index 0000000..278b7f0
--- /dev/null
+++ b/checkpoint-815/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoint-815/tokenizer.json b/checkpoint-815/tokenizer.json
new file mode 100644
index 0000000..1c1d8d5
--- /dev/null
+++ b/checkpoint-815/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/checkpoint-815/tokenizer_config.json b/checkpoint-815/tokenizer_config.json
new file mode 100644
index 0000000..8098161
--- /dev/null
+++ b/checkpoint-815/tokenizer_config.json
@@ -0,0 +1,2064 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "<|end_of_text|>",
+  "tokenizer_class": "PreTrainedTokenizer"
+}
diff --git a/checkpoint-815/trainer_state.json b/checkpoint-815/trainer_state.json
new file mode 100644
index 0000000..a26e97d
--- /dev/null
+++ b/checkpoint-815/trainer_state.json
@@ -0,0 +1,12259 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 815,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001226993865030675,
+      "grad_norm": 151.42932274482283,
+      "learning_rate": 3.2e-08,
+      "logits/chosen": -0.32421875,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -474.0,
+      "logps/rejected": -316.0,
+      "loss": 0.6914,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.00245398773006135,
+      "grad_norm": 162.73344539002164,
+      "learning_rate": 6.4e-08,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -292.0,
+      "loss": 0.6914,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 2
+    },
+    {
+      "epoch": 0.0036809815950920245,
+      "grad_norm": 111.8042794987884,
+      "learning_rate": 9.599999999999999e-08,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -272.0,
+      "loss": 0.6886,
+      "rewards/accuracies": 0.3046875,
+      "rewards/chosen": 0.01531982421875,
+      "rewards/margins": 0.01361083984375,
+      "rewards/rejected": 0.0016632080078125,
+      "step": 3
+    },
+    {
+      "epoch": 0.0049079754601227,
+      "grad_norm": 155.41451533543852,
+      "learning_rate": 1.28e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -242.0,
+      "loss": 0.7131,
+      "rewards/accuracies": 0.328125,
+      "rewards/chosen": -0.00408935546875,
+      "rewards/margins": -0.0218505859375,
+      "rewards/rejected": 0.017822265625,
+      "step": 4
+    },
+    {
+      "epoch": 0.006134969325153374,
+      "grad_norm": 307.019899938903,
+      "learning_rate": 1.6e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -270.0,
+      "loss": 0.6906,
+      "rewards/accuracies": 0.3515625,
+      "rewards/chosen": 0.01171875,
+      "rewards/margins": 0.01251220703125,
+      "rewards/rejected": -0.00078582763671875,
+      "step": 5
+    },
+    {
+      "epoch": 0.007361963190184049,
+      "grad_norm": 340.66722515261114,
+      "learning_rate": 1.9199999999999997e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -306.0,
+      "loss": 0.7151,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.0228271484375,
+      "rewards/margins": -0.034912109375,
+      "rewards/rejected": 0.01220703125,
+      "step": 6
+    },
+    {
+      "epoch": 0.008588957055214725,
+      "grad_norm": 239.24556126143491,
+      "learning_rate": 2.2400000000000002e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -266.0,
+      "loss": 0.7032,
+      "rewards/accuracies": 0.2734375,
+      "rewards/chosen": -0.0003414154052734375,
+      "rewards/margins": -0.01373291015625,
+      "rewards/rejected": 0.01336669921875,
+      "step": 7
+    },
+    {
+      "epoch": 0.0098159509202454,
+      "grad_norm": 139.49365426176607,
+      "learning_rate": 2.56e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -292.0,
+      "loss": 0.7012,
+      "rewards/accuracies": 0.3671875,
+      "rewards/chosen": 0.0284423828125,
+      "rewards/margins": -0.00927734375,
+      "rewards/rejected": 0.037841796875,
+      "step": 8
+    },
+    {
+      "epoch": 0.011042944785276074,
+      "grad_norm": 122.82829055069632,
+      "learning_rate": 2.88e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -294.0,
+      "loss": 0.6871,
+      "rewards/accuracies": 0.375,
+      "rewards/chosen": 0.037353515625,
+      "rewards/margins": 0.0244140625,
+      "rewards/rejected": 0.01300048828125,
+      "step": 9
+    },
+    {
+      "epoch": 0.012269938650306749,
+      "grad_norm": 137.4984961934261,
+      "learning_rate": 3.2e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -278.0,
+      "loss": 0.6605,
+      "rewards/accuracies": 0.46875,
+      "rewards/chosen": 0.057861328125,
+      "rewards/margins": 0.07275390625,
+      "rewards/rejected": -0.01495361328125,
+      "step": 10
+    },
+    {
+      "epoch": 0.013496932515337423,
+      "grad_norm": 154.04829614927058,
+      "learning_rate": 3.52e-07,
+      "logits/chosen": -0.388671875,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -296.0,
+      "loss": 0.6512,
+      "rewards/accuracies": 0.4765625,
+      "rewards/chosen": 0.0625,
+      "rewards/margins": 0.09912109375,
+      "rewards/rejected": -0.036376953125,
+      "step": 11
+    },
+    {
+      "epoch": 0.014723926380368098,
+      "grad_norm": 292.49015304309825,
+      "learning_rate": 3.8399999999999994e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -262.0,
+      "loss": 0.6321,
+      "rewards/accuracies": 0.53125,
+      "rewards/chosen": 0.080078125,
+      "rewards/margins": 0.1435546875,
+      "rewards/rejected": -0.06298828125,
+      "step": 12
+    },
+    {
+      "epoch": 0.015950920245398775,
+      "grad_norm": 179.53722097804638,
+      "learning_rate": 4.16e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -249.0,
+      "loss": 0.6194,
+      "rewards/accuracies": 0.6171875,
+      "rewards/chosen": 0.12109375,
+      "rewards/margins": 0.1845703125,
+      "rewards/rejected": -0.0634765625,
+      "step": 13
+    },
+    {
+      "epoch": 0.01717791411042945,
+      "grad_norm": 108.27195029021392,
+      "learning_rate": 4.4800000000000004e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -472.0,
+      "logps/rejected": -312.0,
+      "loss": 0.5477,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": 0.310546875,
+      "rewards/margins": 0.408203125,
+      "rewards/rejected": -0.09814453125,
+      "step": 14
+    },
+    {
+      "epoch": 0.018404907975460124,
+      "grad_norm": 71.36850002174661,
+      "learning_rate": 4.8e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.18359375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -290.0,
+      "loss": 0.5893,
+      "rewards/accuracies": 0.59375,
+      "rewards/chosen": 0.2412109375,
+      "rewards/margins": 0.337890625,
+      "rewards/rejected": -0.09619140625,
+      "step": 15
+    },
+    {
+      "epoch": 0.0196319018404908,
+      "grad_norm": 80.69275734688873,
+      "learning_rate": 5.12e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -306.0,
+      "loss": 0.5357,
+      "rewards/accuracies": 0.6484375,
+      "rewards/chosen": 0.318359375,
+      "rewards/margins": 0.48828125,
+      "rewards/rejected": -0.1689453125,
+      "step": 16
+    },
+    {
+      "epoch": 0.020858895705521473,
+      "grad_norm": 103.00148200369824,
+      "learning_rate": 5.44e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -255.0,
+      "loss": 0.5279,
+      "rewards/accuracies": 0.6953125,
+      "rewards/chosen": 0.369140625,
+      "rewards/margins": 0.54296875,
+      "rewards/rejected": -0.1728515625,
+      "step": 17
+    },
+    {
+      "epoch": 0.022085889570552148,
+      "grad_norm": 90.51865254706846,
+      "learning_rate": 5.76e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -276.0,
+      "loss": 0.5097,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 0.419921875,
+      "rewards/margins": 0.66015625,
+      "rewards/rejected": -0.23828125,
+      "step": 18
+    },
+    {
+      "epoch": 0.023312883435582823,
+      "grad_norm": 63.171857373704086,
+      "learning_rate": 6.079999999999999e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -290.0,
+      "loss": 0.5237,
+      "rewards/accuracies": 0.6953125,
+      "rewards/chosen": 0.4375,
+      "rewards/margins": 0.80078125,
+      "rewards/rejected": -0.36328125,
+      "step": 19
+    },
+    {
+      "epoch": 0.024539877300613498,
+      "grad_norm": 44.73691218363087,
+      "learning_rate": 6.4e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -462.0,
+      "logps/rejected": -314.0,
+      "loss": 0.4345,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.82421875,
+      "rewards/margins": 1.3984375,
+      "rewards/rejected": -0.5703125,
+      "step": 20
+    },
+    {
+      "epoch": 0.025766871165644172,
+      "grad_norm": 60.19744308670884,
+      "learning_rate": 6.72e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -262.0,
+      "loss": 0.4418,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.62890625,
+      "rewards/margins": 1.1328125,
+      "rewards/rejected": -0.50390625,
+      "step": 21
+    },
+    {
+      "epoch": 0.026993865030674847,
+      "grad_norm": 60.06635545400663,
+      "learning_rate": 7.04e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -294.0,
+      "loss": 0.4638,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 0.640625,
+      "rewards/margins": 1.25,
+      "rewards/rejected": -0.61328125,
+      "step": 22
+    },
+    {
+      "epoch": 0.02822085889570552,
+      "grad_norm": 81.66202093067754,
+      "learning_rate": 7.36e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -306.0,
+      "loss": 0.5093,
+      "rewards/accuracies": 0.7421875,
+      "rewards/chosen": 0.81640625,
+      "rewards/margins": 1.421875,
+      "rewards/rejected": -0.6015625,
+      "step": 23
+    },
+    {
+      "epoch": 0.029447852760736196,
+      "grad_norm": 83.23689816485307,
+      "learning_rate": 7.679999999999999e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -278.0,
+      "loss": 0.511,
+      "rewards/accuracies": 0.7265625,
+      "rewards/chosen": 0.47265625,
+      "rewards/margins": 1.03125,
+      "rewards/rejected": -0.5546875,
+      "step": 24
+    },
+    {
+      "epoch": 0.03067484662576687,
+      "grad_norm": 38.90986069898983,
+      "learning_rate": 8e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.4765625,
+      "logps/chosen": -434.0,
+      "logps/rejected": -300.0,
+      "loss": 0.4067,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.8515625,
+      "rewards/margins": 1.7578125,
+      "rewards/rejected": -0.90625,
+      "step": 25
+    },
+    {
+      "epoch": 0.03190184049079755,
+      "grad_norm": 47.94846803093583,
+      "learning_rate": 7.996694214876033e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -276.0,
+      "loss": 0.4388,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 1.84375,
+      "rewards/rejected": -0.8203125,
+      "step": 26
+    },
+    {
+      "epoch": 0.033128834355828224,
+      "grad_norm": 44.30203970039667,
+      "learning_rate": 7.993388429752066e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -280.0,
+      "loss": 0.3991,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 2.234375,
+      "rewards/rejected": -1.0,
+      "step": 27
+    },
+    {
+      "epoch": 0.0343558282208589,
+      "grad_norm": 38.710283535206564,
+      "learning_rate": 7.990082644628099e-07,
+      "logits/chosen": -0.06591796875,
+      "logits/rejected": -0.1767578125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -262.0,
+      "loss": 0.4113,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 2.015625,
+      "rewards/rejected": -0.74609375,
+      "step": 28
+    },
+    {
+      "epoch": 0.03558282208588957,
+      "grad_norm": 52.910805586215126,
+      "learning_rate": 7.986776859504132e-07,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -322.0,
+      "loss": 0.392,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 2.546875,
+      "rewards/rejected": -1.1015625,
+      "step": 29
+    },
+    {
+      "epoch": 0.03680981595092025,
+      "grad_norm": 45.21313329781316,
+      "learning_rate": 7.983471074380164e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -334.0,
+      "logps/rejected": -266.0,
+      "loss": 0.4322,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 2.03125,
+      "rewards/rejected": -0.93359375,
+      "step": 30
+    },
+    {
+      "epoch": 0.03803680981595092,
+      "grad_norm": 41.258011933772664,
+      "learning_rate": 7.980165289256198e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -276.0,
+      "loss": 0.4115,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 2.765625,
+      "rewards/rejected": -1.1171875,
+      "step": 31
+    },
+    {
+      "epoch": 0.0392638036809816,
+      "grad_norm": 39.16160748489701,
+      "learning_rate": 7.976859504132231e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3836,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5859375,
+      "rewards/margins": 2.71875,
+      "rewards/rejected": -1.1328125,
+      "step": 32
+    },
+    {
+      "epoch": 0.04049079754601227,
+      "grad_norm": 38.84939653302124,
+      "learning_rate": 7.973553719008264e-07,
+      "logits/chosen": -0.0869140625,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -328.0,
+      "logps/rejected": -234.0,
+      "loss": 0.3966,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 2.234375,
+      "rewards/rejected": -1.0859375,
+      "step": 33
+    },
+    {
+      "epoch": 0.04171779141104295,
+      "grad_norm": 44.221302785049176,
+      "learning_rate": 7.970247933884297e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -302.0,
+      "loss": 0.4038,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 2.859375,
+      "rewards/rejected": -1.46875,
+      "step": 34
+    },
+    {
+      "epoch": 0.04294478527607362,
+      "grad_norm": 46.59951208776938,
+      "learning_rate": 7.96694214876033e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -264.0,
+      "loss": 0.3758,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 2.71875,
+      "rewards/rejected": -1.6484375,
+      "step": 35
+    },
+    {
+      "epoch": 0.044171779141104296,
+      "grad_norm": 41.317139252172254,
+      "learning_rate": 7.963636363636364e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3557,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 3.234375,
+      "rewards/rejected": -1.7421875,
+      "step": 36
+    },
+    {
+      "epoch": 0.04539877300613497,
+      "grad_norm": 39.795000825479654,
+      "learning_rate": 7.960330578512397e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3698,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 3.328125,
+      "rewards/rejected": -1.53125,
+      "step": 37
+    },
+    {
+      "epoch": 0.046625766871165646,
+      "grad_norm": 35.75368356475873,
+      "learning_rate": 7.95702479338843e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3582,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 3.6875,
+      "rewards/rejected": -1.671875,
+      "step": 38
+    },
+    {
+      "epoch": 0.04785276073619632,
+      "grad_norm": 46.53573926538421,
+      "learning_rate": 7.953719008264462e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -424.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3835,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 3.671875,
+      "rewards/rejected": -1.984375,
+      "step": 39
+    },
+    {
+      "epoch": 0.049079754601226995,
+      "grad_norm": 33.56284839867345,
+      "learning_rate": 7.950413223140495e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3042,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 3.765625,
+      "rewards/rejected": -2.109375,
+      "step": 40
+    },
+    {
+      "epoch": 0.05030674846625767,
+      "grad_norm": 39.46235435607716,
+      "learning_rate": 7.947107438016528e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -278.0,
+      "loss": 0.3708,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 3.546875,
+      "rewards/rejected": -2.0,
+      "step": 41
+    },
+    {
+      "epoch": 0.051533742331288344,
+      "grad_norm": 37.85600478708204,
+      "learning_rate": 7.943801652892562e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3236,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 3.59375,
+      "rewards/rejected": -2.15625,
+      "step": 42
+    },
+    {
+      "epoch": 0.05276073619631902,
+      "grad_norm": 43.531972474165585,
+      "learning_rate": 7.940495867768595e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3117,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6484375,
+      "rewards/margins": 4.0625,
+      "rewards/rejected": -2.421875,
+      "step": 43
+    },
+    {
+      "epoch": 0.053987730061349694,
+      "grad_norm": 63.54372274745974,
+      "learning_rate": 7.937190082644628e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -300.0,
+      "loss": 0.4265,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 3.625,
+      "rewards/rejected": -2.53125,
+      "step": 44
+    },
+    {
+      "epoch": 0.05521472392638037,
+      "grad_norm": 42.36111410389095,
+      "learning_rate": 7.933884297520661e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3513,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 3.875,
+      "rewards/rejected": -2.640625,
+      "step": 45
+    },
+    {
+      "epoch": 0.05644171779141104,
+      "grad_norm": 44.51420354088873,
+      "learning_rate": 7.930578512396694e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3109,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -2.609375,
+      "step": 46
+    },
+    {
+      "epoch": 0.05766871165644172,
+      "grad_norm": 61.22394416060924,
+      "learning_rate": 7.927272727272727e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -368.0,
+      "loss": 0.4432,
+      "rewards/accuracies": 0.7265625,
+      "rewards/chosen": 1.2109375,
+      "rewards/margins": 3.5,
+      "rewards/rejected": -2.28125,
+      "step": 47
+    },
+    {
+      "epoch": 0.05889570552147239,
+      "grad_norm": 35.37901763029379,
+      "learning_rate": 7.92396694214876e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.361328125,
+      "logps/chosen": -352.0,
+      "logps/rejected": -282.0,
+      "loss": 0.3719,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 3.203125,
+      "rewards/rejected": -2.09375,
+      "step": 48
+    },
+    {
+      "epoch": 0.06012269938650307,
+      "grad_norm": 31.606470718574474,
+      "learning_rate": 7.920661157024793e-07,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2642,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.90625,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.75,
+      "step": 49
+    },
+    {
+      "epoch": 0.06134969325153374,
+      "grad_norm": 48.38174059914219,
+      "learning_rate": 7.917355371900826e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3848,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 3.8125,
+      "rewards/rejected": -2.359375,
+      "step": 50
+    },
+    {
+      "epoch": 0.06257668711656442,
+      "grad_norm": 41.76003291409928,
+      "learning_rate": 7.914049586776859e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -442.0,
+      "logps/rejected": -358.0,
+      "loss": 0.2936,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.890625,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.640625,
+      "step": 51
+    },
+    {
+      "epoch": 0.0638036809815951,
+      "grad_norm": 36.258140724486836,
+      "learning_rate": 7.910743801652892e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3238,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 4.21875,
+      "rewards/rejected": -2.75,
+      "step": 52
+    },
+    {
+      "epoch": 0.06503067484662577,
+      "grad_norm": 38.82725433032811,
+      "learning_rate": 7.907438016528925e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3104,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 4.03125,
+      "rewards/rejected": -2.28125,
+      "step": 53
+    },
+    {
+      "epoch": 0.06625766871165645,
+      "grad_norm": 34.79397739539489,
+      "learning_rate": 7.904132231404959e-07,
+      "logits/chosen": -0.0284423828125,
+      "logits/rejected": -0.208984375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -258.0,
+      "loss": 0.2634,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 3.78125,
+      "rewards/rejected": -2.546875,
+      "step": 54
+    },
+    {
+      "epoch": 0.06748466257668712,
+      "grad_norm": 41.093192913712485,
+      "learning_rate": 7.900826446280992e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3081,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.515625,
+      "step": 55
+    },
+    {
+      "epoch": 0.0687116564417178,
+      "grad_norm": 45.49364963721349,
+      "learning_rate": 7.897520661157025e-07,
+      "logits/chosen": -0.119140625,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3753,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 4.0625,
+      "rewards/rejected": -2.265625,
+      "step": 56
+    },
+    {
+      "epoch": 0.06993865030674846,
+      "grad_norm": 50.666379541833365,
+      "learning_rate": 7.894214876033057e-07,
+      "logits/chosen": -0.1181640625,
+      "logits/rejected": -0.203125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3742,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 3.953125,
+      "rewards/rejected": -2.421875,
+      "step": 57
+    },
+    {
+      "epoch": 0.07116564417177915,
+      "grad_norm": 47.64145593969582,
+      "learning_rate": 7.89090909090909e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.2177734375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3598,
+      "rewards/accuracies": 0.7421875,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 4.125,
+      "rewards/rejected": -2.421875,
+      "step": 58
+    },
+    {
+      "epoch": 0.07239263803680981,
+      "grad_norm": 51.11345473280857,
+      "learning_rate": 7.887603305785123e-07,
+      "logits/chosen": -0.049560546875,
+      "logits/rejected": -0.140625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -292.0,
+      "loss": 0.4133,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 3.28125,
+      "rewards/rejected": -1.9921875,
+      "step": 59
+    },
+    {
+      "epoch": 0.0736196319018405,
+      "grad_norm": 39.0647396184868,
+      "learning_rate": 7.884297520661157e-07,
+      "logits/chosen": -0.12451171875,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -284.0,
+      "loss": 0.3578,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 3.46875,
+      "rewards/rejected": -2.28125,
+      "step": 60
+    },
+    {
+      "epoch": 0.07484662576687116,
+      "grad_norm": 39.12713457741509,
+      "learning_rate": 7.88099173553719e-07,
+      "logits/chosen": -0.125,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3224,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.953125,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.578125,
+      "step": 61
+    },
+    {
+      "epoch": 0.07607361963190185,
+      "grad_norm": 41.544347269900555,
+      "learning_rate": 7.877685950413223e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -428.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3433,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 2.328125,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -2.703125,
+      "step": 62
+    },
+    {
+      "epoch": 0.07730061349693251,
+      "grad_norm": 52.28530938487766,
+      "learning_rate": 7.874380165289256e-07,
+      "logits/chosen": -0.0257568359375,
+      "logits/rejected": -0.197265625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -304.0,
+      "loss": 0.331,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 3.9375,
+      "rewards/rejected": -2.40625,
+      "step": 63
+    },
+    {
+      "epoch": 0.0785276073619632,
+      "grad_norm": 44.343811483698694,
+      "learning_rate": 7.871074380165289e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3628,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -2.5,
+      "step": 64
+    },
+    {
+      "epoch": 0.07975460122699386,
+      "grad_norm": 40.38168553746162,
+      "learning_rate": 7.867768595041322e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3193,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.828125,
+      "rewards/margins": 4.28125,
+      "rewards/rejected": -2.453125,
+      "step": 65
+    },
+    {
+      "epoch": 0.08098159509202454,
+      "grad_norm": 35.90018196646264,
+      "learning_rate": 7.864462809917355e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -318.0,
+      "logps/rejected": -260.0,
+      "loss": 0.2892,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.0,
+      "rewards/margins": 4.8125,
+      "rewards/rejected": -2.8125,
+      "step": 66
+    },
+    {
+      "epoch": 0.08220858895705521,
+      "grad_norm": 39.83708858843296,
+      "learning_rate": 7.861157024793388e-07,
+      "logits/chosen": -0.1005859375,
+      "logits/rejected": -0.224609375,
+      "logps/chosen": -332.0,
+      "logps/rejected": -294.0,
+      "loss": 0.344,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 3.90625,
+      "rewards/rejected": -2.671875,
+      "step": 67
+    },
+    {
+      "epoch": 0.0834355828220859,
+      "grad_norm": 41.25508984701749,
+      "learning_rate": 7.857851239669421e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3597,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 3.953125,
+      "rewards/rejected": -2.46875,
+      "step": 68
+    },
+    {
+      "epoch": 0.08466257668711656,
+      "grad_norm": 35.61462182457114,
+      "learning_rate": 7.854545454545454e-07,
+      "logits/chosen": -0.2451171875,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -282.0,
+      "loss": 0.2797,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.3125,
+      "rewards/margins": 4.96875,
+      "rewards/rejected": -2.65625,
+      "step": 69
+    },
+    {
+      "epoch": 0.08588957055214724,
+      "grad_norm": 43.949799113654095,
+      "learning_rate": 7.851239669421487e-07,
+      "logits/chosen": -0.12109375,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3016,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.328125,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.34375,
+      "step": 70
+    },
+    {
+      "epoch": 0.08711656441717791,
+      "grad_norm": 40.385403541291616,
+      "learning_rate": 7.84793388429752e-07,
+      "logits/chosen": -0.068359375,
+      "logits/rejected": -0.1494140625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -278.0,
+      "loss": 0.3925,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 3.625,
+      "rewards/rejected": -2.265625,
+      "step": 71
+    },
+    {
+      "epoch": 0.08834355828220859,
+      "grad_norm": 42.66471807743761,
+      "learning_rate": 7.844628099173554e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -430.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2955,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9609375,
+      "rewards/margins": 4.25,
+      "rewards/rejected": -2.28125,
+      "step": 72
+    },
+    {
+      "epoch": 0.08957055214723926,
+      "grad_norm": 45.57465316917817,
+      "learning_rate": 7.841322314049587e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -348.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3924,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.5625,
+      "step": 73
+    },
+    {
+      "epoch": 0.09079754601226994,
+      "grad_norm": 41.41678309244041,
+      "learning_rate": 7.838016528925619e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -354.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2911,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 3.9375,
+      "rewards/rejected": -2.65625,
+      "step": 74
+    },
+    {
+      "epoch": 0.09202453987730061,
+      "grad_norm": 36.36740264063768,
+      "learning_rate": 7.834710743801652e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -318.0,
+      "loss": 0.271,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.703125,
+      "step": 75
+    },
+    {
+      "epoch": 0.09325153374233129,
+      "grad_norm": 51.152036733559505,
+      "learning_rate": 7.831404958677685e-07,
+      "logits/chosen": -0.1513671875,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -288.0,
+      "loss": 0.3262,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -2.53125,
+      "step": 76
+    },
+    {
+      "epoch": 0.09447852760736196,
+      "grad_norm": 34.55182641721203,
+      "learning_rate": 7.828099173553718e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2774,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.4453125,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -2.875,
+      "step": 77
+    },
+    {
+      "epoch": 0.09570552147239264,
+      "grad_norm": 40.758007342891354,
+      "learning_rate": 7.824793388429752e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2953,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -2.859375,
+      "step": 78
+    },
+    {
+      "epoch": 0.09693251533742331,
+      "grad_norm": 40.243771187807226,
+      "learning_rate": 7.821487603305785e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -286.0,
+      "loss": 0.324,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.796875,
+      "step": 79
+    },
+    {
+      "epoch": 0.09815950920245399,
+      "grad_norm": 37.86169713828253,
+      "learning_rate": 7.818181818181818e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2795,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.0,
+      "step": 80
+    },
+    {
+      "epoch": 0.09938650306748466,
+      "grad_norm": 36.69542578194377,
+      "learning_rate": 7.814876033057851e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -436.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.9375,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.09375,
+      "step": 81
+    },
+    {
+      "epoch": 0.10061349693251534,
+      "grad_norm": 52.318975452049024,
+      "learning_rate": 7.811570247933884e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3562,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -3.1875,
+      "step": 82
+    },
+    {
+      "epoch": 0.10184049079754601,
+      "grad_norm": 48.36250064527517,
+      "learning_rate": 7.808264462809916e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3065,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 4.5625,
+      "rewards/rejected": -3.234375,
+      "step": 83
+    },
+    {
+      "epoch": 0.10306748466257669,
+      "grad_norm": 56.4532578365936,
+      "learning_rate": 7.80495867768595e-07,
+      "logits/chosen": -0.11083984375,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -448.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2942,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -3.53125,
+      "step": 84
+    },
+    {
+      "epoch": 0.10429447852760736,
+      "grad_norm": 40.84421229525563,
+      "learning_rate": 7.801652892561983e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3021,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.484375,
+      "step": 85
+    },
+    {
+      "epoch": 0.10552147239263804,
+      "grad_norm": 42.8824866796569,
+      "learning_rate": 7.798347107438016e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -302.0,
+      "loss": 0.3346,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.03125,
+      "step": 86
+    },
+    {
+      "epoch": 0.1067484662576687,
+      "grad_norm": 32.177789360920215,
+      "learning_rate": 7.795041322314049e-07,
+      "logits/chosen": -0.17578125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -260.0,
+      "loss": 0.2835,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 4.25,
+      "rewards/rejected": -3.46875,
+      "step": 87
+    },
+    {
+      "epoch": 0.10797546012269939,
+      "grad_norm": 34.28909140331474,
+      "learning_rate": 7.791735537190082e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -276.0,
+      "loss": 0.265,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -3.3125,
+      "step": 88
+    },
+    {
+      "epoch": 0.10920245398773006,
+      "grad_norm": 42.40518360379319,
+      "learning_rate": 7.788429752066116e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3464,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.4140625,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -3.09375,
+      "step": 89
+    },
+    {
+      "epoch": 0.11042944785276074,
+      "grad_norm": 49.53932371851528,
+      "learning_rate": 7.785123966942149e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3253,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.98828125,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -3.3125,
+      "step": 90
+    },
+    {
+      "epoch": 0.1116564417177914,
+      "grad_norm": 40.48862678893552,
+      "learning_rate": 7.781818181818182e-07,
+      "logits/chosen": -0.039794921875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2845,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -3.203125,
+      "step": 91
+    },
+    {
+      "epoch": 0.11288343558282209,
+      "grad_norm": 44.87681376045074,
+      "learning_rate": 7.778512396694214e-07,
+      "logits/chosen": -0.126953125,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2738,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.5,
+      "step": 92
+    },
+    {
+      "epoch": 0.11411042944785275,
+      "grad_norm": 36.61224632474363,
+      "learning_rate": 7.775206611570247e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -300.0,
+      "loss": 0.282,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.3125,
+      "step": 93
+    },
+    {
+      "epoch": 0.11533742331288344,
+      "grad_norm": 42.24461574764279,
+      "learning_rate": 7.77190082644628e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -432.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2813,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.234375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.140625,
+      "step": 94
+    },
+    {
+      "epoch": 0.1165644171779141,
+      "grad_norm": 42.13498046583409,
+      "learning_rate": 7.768595041322314e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.7734375,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -3.390625,
+      "step": 95
+    },
+    {
+      "epoch": 0.11779141104294479,
+      "grad_norm": 39.13305361039425,
+      "learning_rate": 7.765289256198347e-07,
+      "logits/chosen": -0.2333984375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -310.0,
+      "loss": 0.253,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.203125,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -3.40625,
+      "step": 96
+    },
+    {
+      "epoch": 0.11901840490797547,
+      "grad_norm": 46.1275059887527,
+      "learning_rate": 7.76198347107438e-07,
+      "logits/chosen": -0.2294921875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -426.0,
+      "logps/rejected": -352.0,
+      "loss": 0.357,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -2.890625,
+      "step": 97
+    },
+    {
+      "epoch": 0.12024539877300613,
+      "grad_norm": 38.580559184520915,
+      "learning_rate": 7.758677685950413e-07,
+      "logits/chosen": -0.041015625,
+      "logits/rejected": -0.248046875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -282.0,
+      "loss": 0.3015,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.828125,
+      "step": 98
+    },
+    {
+      "epoch": 0.12147239263803682,
+      "grad_norm": 38.975117366609354,
+      "learning_rate": 7.755371900826446e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -340.0,
+      "logps/rejected": -244.0,
+      "loss": 0.2586,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 4.09375,
+      "rewards/rejected": -2.84375,
+      "step": 99
+    },
+    {
+      "epoch": 0.12269938650306748,
+      "grad_norm": 38.2314125667712,
+      "learning_rate": 7.752066115702479e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3048,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 4.875,
+      "rewards/rejected": -3.09375,
+      "step": 100
+    },
+    {
+      "epoch": 0.12392638036809817,
+      "grad_norm": 47.72088602593496,
+      "learning_rate": 7.748760330578512e-07,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3549,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.921875,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -2.984375,
+      "step": 101
+    },
+    {
+      "epoch": 0.12515337423312883,
+      "grad_norm": 39.20874470259191,
+      "learning_rate": 7.745454545454545e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3073,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.9375,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -2.71875,
+      "step": 102
+    },
+    {
+      "epoch": 0.1263803680981595,
+      "grad_norm": 36.963577318972284,
+      "learning_rate": 7.742148760330578e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2878,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7265625,
+      "rewards/margins": 4.46875,
+      "rewards/rejected": -2.75,
+      "step": 103
+    },
+    {
+      "epoch": 0.1276073619631902,
+      "grad_norm": 50.89003123288155,
+      "learning_rate": 7.738842975206611e-07,
+      "logits/chosen": -0.107421875,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -414.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3679,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 4.40625,
+      "rewards/rejected": -2.71875,
+      "step": 104
+    },
+    {
+      "epoch": 0.12883435582822086,
+      "grad_norm": 34.19259318441508,
+      "learning_rate": 7.735537190082644e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2586,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.984375,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.140625,
+      "step": 105
+    },
+    {
+      "epoch": 0.13006134969325153,
+      "grad_norm": 38.48374397973381,
+      "learning_rate": 7.732231404958677e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3111,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.9140625,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -2.828125,
+      "step": 106
+    },
+    {
+      "epoch": 0.1312883435582822,
+      "grad_norm": 41.75637341421824,
+      "learning_rate": 7.728925619834711e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2869,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 4.53125,
+      "rewards/rejected": -2.84375,
+      "step": 107
+    },
+    {
+      "epoch": 0.1325153374233129,
+      "grad_norm": 33.954750583294235,
+      "learning_rate": 7.725619834710744e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2625,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.328125,
+      "step": 108
+    },
+    {
+      "epoch": 0.13374233128834356,
+      "grad_norm": 36.44619320062165,
+      "learning_rate": 7.722314049586777e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2958,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.6875,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -2.96875,
+      "step": 109
+    },
+    {
+      "epoch": 0.13496932515337423,
+      "grad_norm": 41.465536696868014,
+      "learning_rate": 7.719008264462809e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3213,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 2.234375,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.015625,
+      "step": 110
+    },
+    {
+      "epoch": 0.1361963190184049,
+      "grad_norm": 34.97820982991418,
+      "learning_rate": 7.715702479338842e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -330.0,
+      "logps/rejected": -268.0,
+      "loss": 0.2777,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 4.40625,
+      "rewards/rejected": -2.75,
+      "step": 111
+    },
+    {
+      "epoch": 0.1374233128834356,
+      "grad_norm": 37.33233370442031,
+      "learning_rate": 7.712396694214875e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -340.0,
+      "logps/rejected": -278.0,
+      "loss": 0.2729,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 4.6875,
+      "rewards/rejected": -3.125,
+      "step": 112
+    },
+    {
+      "epoch": 0.13865030674846626,
+      "grad_norm": 39.32831836749558,
+      "learning_rate": 7.709090909090909e-07,
+      "logits/chosen": -0.1044921875,
+      "logits/rejected": -0.2001953125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -302.0,
+      "loss": 0.3238,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.84375,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -2.890625,
+      "step": 113
+    },
+    {
+      "epoch": 0.13987730061349693,
+      "grad_norm": 43.77160490529073,
+      "learning_rate": 7.705785123966942e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3356,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -2.96875,
+      "step": 114
+    },
+    {
+      "epoch": 0.1411042944785276,
+      "grad_norm": 45.86173030028754,
+      "learning_rate": 7.702479338842975e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3666,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 4.0,
+      "rewards/rejected": -2.96875,
+      "step": 115
+    },
+    {
+      "epoch": 0.1423312883435583,
+      "grad_norm": 31.833834890400677,
+      "learning_rate": 7.699173553719008e-07,
+      "logits/chosen": -0.2001953125,
+      "logits/rejected": -0.439453125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2561,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.078125,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.140625,
+      "step": 116
+    },
+    {
+      "epoch": 0.14355828220858896,
+      "grad_norm": 34.65342816309822,
+      "learning_rate": 7.695867768595041e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -324.0,
+      "logps/rejected": -262.0,
+      "loss": 0.2802,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -3.453125,
+      "step": 117
+    },
+    {
+      "epoch": 0.14478527607361963,
+      "grad_norm": 40.64680750199124,
+      "learning_rate": 7.692561983471075e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3232,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.09375,
+      "step": 118
+    },
+    {
+      "epoch": 0.1460122699386503,
+      "grad_norm": 39.06542225750342,
+      "learning_rate": 7.689256198347107e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3009,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -3.34375,
+      "step": 119
+    },
+    {
+      "epoch": 0.147239263803681,
+      "grad_norm": 45.793161618342445,
+      "learning_rate": 7.68595041322314e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2967,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 2.421875,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -3.234375,
+      "step": 120
+    },
+    {
+      "epoch": 0.14846625766871166,
+      "grad_norm": 36.35506762175528,
+      "learning_rate": 7.682644628099173e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2827,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 2.265625,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -3.09375,
+      "step": 121
+    },
+    {
+      "epoch": 0.14969325153374233,
+      "grad_norm": 43.61172830028863,
+      "learning_rate": 7.679338842975206e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2954,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -3.015625,
+      "step": 122
+    },
+    {
+      "epoch": 0.150920245398773,
+      "grad_norm": 44.76510892093205,
+      "learning_rate": 7.676033057851239e-07,
+      "logits/chosen": -0.054443359375,
+      "logits/rejected": -0.2294921875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -278.0,
+      "loss": 0.3266,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 4.40625,
+      "rewards/rejected": -3.171875,
+      "step": 123
+    },
+    {
+      "epoch": 0.1521472392638037,
+      "grad_norm": 34.74558330134135,
+      "learning_rate": 7.672727272727273e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -350.0,
+      "logps/rejected": -280.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -3.46875,
+      "step": 124
+    },
+    {
+      "epoch": 0.15337423312883436,
+      "grad_norm": 42.40382525839625,
+      "learning_rate": 7.669421487603306e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2816,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.765625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.5625,
+      "step": 125
+    },
+    {
+      "epoch": 0.15460122699386503,
+      "grad_norm": 54.09847924655591,
+      "learning_rate": 7.666115702479339e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -302.0,
+      "loss": 0.3503,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 4.34375,
+      "rewards/rejected": -3.34375,
+      "step": 126
+    },
+    {
+      "epoch": 0.1558282208588957,
+      "grad_norm": 42.126209906017394,
+      "learning_rate": 7.662809917355372e-07,
+      "logits/chosen": -0.09375,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3524,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.96875,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.15625,
+      "step": 127
+    },
+    {
+      "epoch": 0.1570552147239264,
+      "grad_norm": 41.454801423663085,
+      "learning_rate": 7.659504132231404e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2841,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8671875,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.390625,
+      "step": 128
+    },
+    {
+      "epoch": 0.15828220858895706,
+      "grad_norm": 37.48031578456685,
+      "learning_rate": 7.656198347107437e-07,
+      "logits/chosen": -0.0478515625,
+      "logits/rejected": -0.1796875,
+      "logps/chosen": -336.0,
+      "logps/rejected": -284.0,
+      "loss": 0.3245,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 4.3125,
+      "rewards/rejected": -3.0,
+      "step": 129
+    },
+    {
+      "epoch": 0.15950920245398773,
+      "grad_norm": 38.14450880210714,
+      "learning_rate": 7.65289256198347e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3273,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -3.421875,
+      "step": 130
+    },
+    {
+      "epoch": 0.1607361963190184,
+      "grad_norm": 35.19502262660919,
+      "learning_rate": 7.649586776859504e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2106,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.953125,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.59375,
+      "step": 131
+    },
+    {
+      "epoch": 0.1619631901840491,
+      "grad_norm": 42.07184086801236,
+      "learning_rate": 7.646280991735537e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3252,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.765625,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -3.40625,
+      "step": 132
+    },
+    {
+      "epoch": 0.16319018404907976,
+      "grad_norm": 43.059635351413924,
+      "learning_rate": 7.64297520661157e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2872,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.703125,
+      "step": 133
+    },
+    {
+      "epoch": 0.16441717791411042,
+      "grad_norm": 48.09563853008669,
+      "learning_rate": 7.639669421487603e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -450.0,
+      "logps/rejected": -358.0,
+      "loss": 0.2946,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 4.625,
+      "rewards/rejected": -3.109375,
+      "step": 134
+    },
+    {
+      "epoch": 0.1656441717791411,
+      "grad_norm": 46.36508764131004,
+      "learning_rate": 7.636363636363636e-07,
+      "logits/chosen": -0.05712890625,
+      "logits/rejected": -0.171875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3654,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.9765625,
+      "rewards/margins": 4.375,
+      "rewards/rejected": -3.390625,
+      "step": 135
+    },
+    {
+      "epoch": 0.1668711656441718,
+      "grad_norm": 39.349987951574626,
+      "learning_rate": 7.63305785123967e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3195,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -3.78125,
+      "step": 136
+    },
+    {
+      "epoch": 0.16809815950920245,
+      "grad_norm": 46.78307320347251,
+      "learning_rate": 7.629752066115702e-07,
+      "logits/chosen": -0.061279296875,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3621,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 4.25,
+      "rewards/rejected": -3.078125,
+      "step": 137
+    },
+    {
+      "epoch": 0.16932515337423312,
+      "grad_norm": 37.18351460769713,
+      "learning_rate": 7.626446280991735e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3022,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.46875,
+      "step": 138
+    },
+    {
+      "epoch": 0.1705521472392638,
+      "grad_norm": 42.545002180919454,
+      "learning_rate": 7.623140495867768e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -412.0,
+      "logps/rejected": -362.0,
+      "loss": 0.352,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -3.828125,
+      "step": 139
+    },
+    {
+      "epoch": 0.17177914110429449,
+      "grad_norm": 47.45496980791103,
+      "learning_rate": 7.619834710743801e-07,
+      "logits/chosen": -0.10400390625,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3166,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.5625,
+      "step": 140
+    },
+    {
+      "epoch": 0.17300613496932515,
+      "grad_norm": 34.622965144928166,
+      "learning_rate": 7.616528925619834e-07,
+      "logits/chosen": -0.0849609375,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.251,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.515625,
+      "step": 141
+    },
+    {
+      "epoch": 0.17423312883435582,
+      "grad_norm": 40.37071693879945,
+      "learning_rate": 7.613223140495868e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3233,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.640625,
+      "step": 142
+    },
+    {
+      "epoch": 0.1754601226993865,
+      "grad_norm": 41.48963188375289,
+      "learning_rate": 7.609917355371901e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2941,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.828125,
+      "step": 143
+    },
+    {
+      "epoch": 0.17668711656441718,
+      "grad_norm": 40.697834722668915,
+      "learning_rate": 7.606611570247934e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2905,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.96875,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.59375,
+      "step": 144
+    },
+    {
+      "epoch": 0.17791411042944785,
+      "grad_norm": 39.56968971060715,
+      "learning_rate": 7.603305785123967e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2715,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.65625,
+      "step": 145
+    },
+    {
+      "epoch": 0.17914110429447852,
+      "grad_norm": 34.52865223755899,
+      "learning_rate": 7.599999999999999e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2436,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.78125,
+      "step": 146
+    },
+    {
+      "epoch": 0.18036809815950922,
+      "grad_norm": 35.86987627703754,
+      "learning_rate": 7.596694214876032e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2741,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.84375,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.71875,
+      "step": 147
+    },
+    {
+      "epoch": 0.18159509202453988,
+      "grad_norm": 49.681572441729514,
+      "learning_rate": 7.593388429752066e-07,
+      "logits/chosen": -0.23828125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -412.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3761,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.625,
+      "step": 148
+    },
+    {
+      "epoch": 0.18282208588957055,
+      "grad_norm": 43.932202256712266,
+      "learning_rate": 7.590082644628099e-07,
+      "logits/chosen": -0.11474609375,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2534,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.984375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.390625,
+      "step": 149
+    },
+    {
+      "epoch": 0.18404907975460122,
+      "grad_norm": 39.402366627439726,
+      "learning_rate": 7.586776859504132e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3158,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 4.71875,
+      "rewards/rejected": -3.09375,
+      "step": 150
+    },
+    {
+      "epoch": 0.18527607361963191,
+      "grad_norm": 41.79252221918161,
+      "learning_rate": 7.583471074380165e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3211,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -3.03125,
+      "step": 151
+    },
+    {
+      "epoch": 0.18650306748466258,
+      "grad_norm": 35.715508495832665,
+      "learning_rate": 7.580165289256198e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -298.0,
+      "loss": 0.2622,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.921875,
+      "step": 152
+    },
+    {
+      "epoch": 0.18773006134969325,
+      "grad_norm": 36.417969122212845,
+      "learning_rate": 7.576859504132231e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2871,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.625,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.390625,
+      "step": 153
+    },
+    {
+      "epoch": 0.18895705521472392,
+      "grad_norm": 45.428337431051204,
+      "learning_rate": 7.573553719008265e-07,
+      "logits/chosen": -0.09716796875,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3165,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 2.046875,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -2.984375,
+      "step": 154
+    },
+    {
+      "epoch": 0.1901840490797546,
+      "grad_norm": 28.34310543484037,
+      "learning_rate": 7.570247933884297e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1864,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 2.5625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -3.65625,
+      "step": 155
+    },
+    {
+      "epoch": 0.19141104294478528,
+      "grad_norm": 43.05423041815277,
+      "learning_rate": 7.56694214876033e-07,
+      "logits/chosen": -0.11865234375,
+      "logits/rejected": -0.2314453125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2995,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 4.5625,
+      "rewards/rejected": -3.09375,
+      "step": 156
+    },
+    {
+      "epoch": 0.19263803680981595,
+      "grad_norm": 35.31331641855709,
+      "learning_rate": 7.563636363636363e-07,
+      "logits/chosen": -0.11474609375,
+      "logits/rejected": -0.244140625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2791,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 4.875,
+      "rewards/rejected": -3.328125,
+      "step": 157
+    },
+    {
+      "epoch": 0.19386503067484662,
+      "grad_norm": 32.52885612180771,
+      "learning_rate": 7.560330578512396e-07,
+      "logits/chosen": -0.30859375,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2615,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.6875,
+      "step": 158
+    },
+    {
+      "epoch": 0.1950920245398773,
+      "grad_norm": 40.74250145862389,
+      "learning_rate": 7.55702479338843e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -318.0,
+      "loss": 0.312,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.25,
+      "step": 159
+    },
+    {
+      "epoch": 0.19631901840490798,
+      "grad_norm": 32.03909927740134,
+      "learning_rate": 7.553719008264463e-07,
+      "logits/chosen": -0.10009765625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2149,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.765625,
+      "step": 160
+    },
+    {
+      "epoch": 0.19754601226993865,
+      "grad_norm": 33.48229976640365,
+      "learning_rate": 7.550413223140496e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2313,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.8984375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -3.703125,
+      "step": 161
+    },
+    {
+      "epoch": 0.19877300613496932,
+      "grad_norm": 44.9989690142436,
+      "learning_rate": 7.547107438016529e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3547,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.390625,
+      "step": 162
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 45.392211600179095,
+      "learning_rate": 7.543801652892561e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3203,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.4375,
+      "step": 163
+    },
+    {
+      "epoch": 0.20122699386503068,
+      "grad_norm": 30.10144812491185,
+      "learning_rate": 7.540495867768594e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.515625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -298.0,
+      "loss": 0.2205,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -3.953125,
+      "step": 164
+    },
+    {
+      "epoch": 0.20245398773006135,
+      "grad_norm": 41.70055038557836,
+      "learning_rate": 7.537190082644627e-07,
+      "logits/chosen": -0.078125,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -440.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2684,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.7578125,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.484375,
+      "step": 165
+    },
+    {
+      "epoch": 0.20368098159509201,
+      "grad_norm": 40.19134249676154,
+      "learning_rate": 7.533884297520661e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3382,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -3.296875,
+      "step": 166
+    },
+    {
+      "epoch": 0.2049079754601227,
+      "grad_norm": 40.73221481198693,
+      "learning_rate": 7.530578512396694e-07,
+      "logits/chosen": -0.1240234375,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3475,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 2.21875,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -3.40625,
+      "step": 167
+    },
+    {
+      "epoch": 0.20613496932515338,
+      "grad_norm": 36.56704570459498,
+      "learning_rate": 7.527272727272727e-07,
+      "logits/chosen": -0.10888671875,
+      "logits/rejected": -0.2236328125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -276.0,
+      "loss": 0.2594,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.0,
+      "step": 168
+    },
+    {
+      "epoch": 0.20736196319018405,
+      "grad_norm": 36.86789794879476,
+      "learning_rate": 7.52396694214876e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2793,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.75,
+      "step": 169
+    },
+    {
+      "epoch": 0.2085889570552147,
+      "grad_norm": 43.700977925857806,
+      "learning_rate": 7.520661157024793e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -448.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2994,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.046875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -3.765625,
+      "step": 170
+    },
+    {
+      "epoch": 0.2098159509202454,
+      "grad_norm": 33.49786387415663,
+      "learning_rate": 7.517355371900827e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2804,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.7734375,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.0625,
+      "step": 171
+    },
+    {
+      "epoch": 0.21104294478527608,
+      "grad_norm": 36.775839118353844,
+      "learning_rate": 7.514049586776859e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -436.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2514,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.03125,
+      "step": 172
+    },
+    {
+      "epoch": 0.21226993865030674,
+      "grad_norm": 29.010747243935523,
+      "learning_rate": 7.510743801652892e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -422.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2075,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.296875,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.125,
+      "step": 173
+    },
+    {
+      "epoch": 0.2134969325153374,
+      "grad_norm": 42.10282379400345,
+      "learning_rate": 7.507438016528925e-07,
+      "logits/chosen": -0.035400390625,
+      "logits/rejected": -0.248046875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3371,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.890625,
+      "step": 174
+    },
+    {
+      "epoch": 0.2147239263803681,
+      "grad_norm": 44.4053310320228,
+      "learning_rate": 7.504132231404958e-07,
+      "logits/chosen": -0.17578125,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3132,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.9609375,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -3.921875,
+      "step": 175
+    },
+    {
+      "epoch": 0.21595092024539878,
+      "grad_norm": 38.79125305740806,
+      "learning_rate": 7.500826446280991e-07,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -428.0,
+      "logps/rejected": -354.0,
+      "loss": 0.315,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -3.875,
+      "step": 176
+    },
+    {
+      "epoch": 0.21717791411042944,
+      "grad_norm": 29.08349894872895,
+      "learning_rate": 7.497520661157025e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2249,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.453125,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -4.375,
+      "step": 177
+    },
+    {
+      "epoch": 0.2184049079754601,
+      "grad_norm": 39.84454997153339,
+      "learning_rate": 7.494214876033058e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3147,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.375,
+      "step": 178
+    },
+    {
+      "epoch": 0.2196319018404908,
+      "grad_norm": 32.49552053062871,
+      "learning_rate": 7.490909090909091e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2585,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.34375,
+      "step": 179
+    },
+    {
+      "epoch": 0.22085889570552147,
+      "grad_norm": 41.53796098392134,
+      "learning_rate": 7.487603305785124e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3355,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.75,
+      "step": 180
+    },
+    {
+      "epoch": 0.22208588957055214,
+      "grad_norm": 43.27113784190138,
+      "learning_rate": 7.484297520661156e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -372.0,
+      "loss": 0.3558,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.03125,
+      "step": 181
+    },
+    {
+      "epoch": 0.2233128834355828,
+      "grad_norm": 40.060092182372635,
+      "learning_rate": 7.480991735537189e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3329,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.92578125,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -4.0625,
+      "step": 182
+    },
+    {
+      "epoch": 0.2245398773006135,
+      "grad_norm": 42.893284410160845,
+      "learning_rate": 7.477685950413223e-07,
+      "logits/chosen": -0.0888671875,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -374.0,
+      "loss": 0.3303,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -3.703125,
+      "step": 183
+    },
+    {
+      "epoch": 0.22576687116564417,
+      "grad_norm": 44.81738203832481,
+      "learning_rate": 7.474380165289256e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.2431640625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3382,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.0625,
+      "step": 184
+    },
+    {
+      "epoch": 0.22699386503067484,
+      "grad_norm": 49.18179746490089,
+      "learning_rate": 7.471074380165289e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3425,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -3.734375,
+      "step": 185
+    },
+    {
+      "epoch": 0.2282208588957055,
+      "grad_norm": 38.62632364463619,
+      "learning_rate": 7.467768595041322e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3061,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.953125,
+      "step": 186
+    },
+    {
+      "epoch": 0.2294478527607362,
+      "grad_norm": 37.48238102625356,
+      "learning_rate": 7.464462809917355e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2719,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.375,
+      "step": 187
+    },
+    {
+      "epoch": 0.23067484662576687,
+      "grad_norm": 33.9965710800535,
+      "learning_rate": 7.461157024793388e-07,
+      "logits/chosen": -0.0712890625,
+      "logits/rejected": -0.1689453125,
+      "logps/chosen": -332.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2908,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.09375,
+      "step": 188
+    },
+    {
+      "epoch": 0.23190184049079754,
+      "grad_norm": 32.6251624460159,
+      "learning_rate": 7.457851239669422e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2196,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.3125,
+      "step": 189
+    },
+    {
+      "epoch": 0.2331288343558282,
+      "grad_norm": 33.86427029616883,
+      "learning_rate": 7.454545454545454e-07,
+      "logits/chosen": -0.05810546875,
+      "logits/rejected": -0.224609375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2859,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.875,
+      "step": 190
+    },
+    {
+      "epoch": 0.2343558282208589,
+      "grad_norm": 48.95754374693823,
+      "learning_rate": 7.451239669421487e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3495,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.1875,
+      "step": 191
+    },
+    {
+      "epoch": 0.23558282208588957,
+      "grad_norm": 44.694192367882906,
+      "learning_rate": 7.44793388429752e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -292.0,
+      "loss": 0.3014,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -3.578125,
+      "step": 192
+    },
+    {
+      "epoch": 0.23680981595092024,
+      "grad_norm": 34.25684662561766,
+      "learning_rate": 7.444628099173553e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2177,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.28125,
+      "step": 193
+    },
+    {
+      "epoch": 0.23803680981595093,
+      "grad_norm": 39.119509435238996,
+      "learning_rate": 7.441322314049586e-07,
+      "logits/chosen": -0.306640625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2953,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.4375,
+      "step": 194
+    },
+    {
+      "epoch": 0.2392638036809816,
+      "grad_norm": 34.817335787332766,
+      "learning_rate": 7.43801652892562e-07,
+      "logits/chosen": -0.1181640625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2301,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.4375,
+      "step": 195
+    },
+    {
+      "epoch": 0.24049079754601227,
+      "grad_norm": 40.102802913255154,
+      "learning_rate": 7.434710743801653e-07,
+      "logits/chosen": -0.09130859375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -290.0,
+      "loss": 0.297,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -3.828125,
+      "step": 196
+    },
+    {
+      "epoch": 0.24171779141104294,
+      "grad_norm": 34.798967486094014,
+      "learning_rate": 7.431404958677686e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2803,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.0625,
+      "step": 197
+    },
+    {
+      "epoch": 0.24294478527607363,
+      "grad_norm": 39.54376141056963,
+      "learning_rate": 7.428099173553719e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.2265625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -258.0,
+      "loss": 0.3347,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.671875,
+      "step": 198
+    },
+    {
+      "epoch": 0.2441717791411043,
+      "grad_norm": 29.961664099131877,
+      "learning_rate": 7.424793388429751e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -428.0,
+      "logps/rejected": -310.0,
+      "loss": 0.1903,
+      "rewards/accuracies": 0.9140625,
+      "rewards/chosen": 2.359375,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.09375,
+      "step": 199
+    },
+    {
+      "epoch": 0.24539877300613497,
+      "grad_norm": 41.86791512110987,
+      "learning_rate": 7.421487603305784e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -370.0,
+      "loss": 0.3056,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.203125,
+      "rewards/margins": 6.125,
+      "rewards/rejected": -3.921875,
+      "step": 200
+    },
+    {
+      "epoch": 0.24662576687116564,
+      "grad_norm": 32.0817049971399,
+      "learning_rate": 7.418181818181818e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2166,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 2.0625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -3.734375,
+      "step": 201
+    },
+    {
+      "epoch": 0.24785276073619633,
+      "grad_norm": 35.7245529395789,
+      "learning_rate": 7.414876033057851e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2641,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -3.640625,
+      "step": 202
+    },
+    {
+      "epoch": 0.249079754601227,
+      "grad_norm": 42.27175209208064,
+      "learning_rate": 7.411570247933884e-07,
+      "logits/chosen": -0.08935546875,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4055,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8828125,
+      "rewards/margins": 4.90625,
+      "rewards/rejected": -3.03125,
+      "step": 203
+    },
+    {
+      "epoch": 0.25030674846625767,
+      "grad_norm": 39.50994905620242,
+      "learning_rate": 7.408264462809917e-07,
+      "logits/chosen": -0.0771484375,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -268.0,
+      "loss": 0.3395,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -3.671875,
+      "step": 204
+    },
+    {
+      "epoch": 0.25153374233128833,
+      "grad_norm": 37.49332179023069,
+      "learning_rate": 7.40495867768595e-07,
+      "logits/chosen": -0.02880859375,
+      "logits/rejected": -0.173828125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3209,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.84375,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -3.265625,
+      "step": 205
+    },
+    {
+      "epoch": 0.252760736196319,
+      "grad_norm": 27.399492756017406,
+      "learning_rate": 7.401652892561984e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2051,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.28125,
+      "step": 206
+    },
+    {
+      "epoch": 0.25398773006134967,
+      "grad_norm": 43.41671912422527,
+      "learning_rate": 7.398347107438017e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3584,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 2.140625,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -3.71875,
+      "step": 207
+    },
+    {
+      "epoch": 0.2552147239263804,
+      "grad_norm": 38.06892744188352,
+      "learning_rate": 7.395041322314049e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3179,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.71875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.09375,
+      "step": 208
+    },
+    {
+      "epoch": 0.25644171779141106,
+      "grad_norm": 38.21598347000612,
+      "learning_rate": 7.391735537190082e-07,
+      "logits/chosen": -0.2177734375,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -336.0,
+      "loss": 0.306,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -3.90625,
+      "step": 209
+    },
+    {
+      "epoch": 0.25766871165644173,
+      "grad_norm": 45.46016539899385,
+      "learning_rate": 7.388429752066115e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -364.0,
+      "loss": 0.3476,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.9375,
+      "step": 210
+    },
+    {
+      "epoch": 0.2588957055214724,
+      "grad_norm": 38.3264065355804,
+      "learning_rate": 7.385123966942148e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3035,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.53125,
+      "step": 211
+    },
+    {
+      "epoch": 0.26012269938650306,
+      "grad_norm": 37.36895578645757,
+      "learning_rate": 7.381818181818182e-07,
+      "logits/chosen": -0.1904296875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2805,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.28125,
+      "step": 212
+    },
+    {
+      "epoch": 0.26134969325153373,
+      "grad_norm": 33.79330367194955,
+      "learning_rate": 7.378512396694215e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -284.0,
+      "loss": 0.2588,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.9375,
+      "step": 213
+    },
+    {
+      "epoch": 0.2625766871165644,
+      "grad_norm": 41.91194027673872,
+      "learning_rate": 7.375206611570248e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3085,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5703125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -4.125,
+      "step": 214
+    },
+    {
+      "epoch": 0.26380368098159507,
+      "grad_norm": 42.67216156371333,
+      "learning_rate": 7.371900826446281e-07,
+      "logits/chosen": -0.0927734375,
+      "logits/rejected": -0.1953125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3066,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 4.96875,
+      "rewards/rejected": -3.8125,
+      "step": 215
+    },
+    {
+      "epoch": 0.2650306748466258,
+      "grad_norm": 41.6917127385005,
+      "learning_rate": 7.368595041322314e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -330.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3077,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 4.65625,
+      "rewards/rejected": -3.734375,
+      "step": 216
+    },
+    {
+      "epoch": 0.26625766871165646,
+      "grad_norm": 41.999916191643734,
+      "learning_rate": 7.365289256198346e-07,
+      "logits/chosen": -0.05322265625,
+      "logits/rejected": -0.20703125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3621,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.421875,
+      "step": 217
+    },
+    {
+      "epoch": 0.2674846625766871,
+      "grad_norm": 48.35045291177634,
+      "learning_rate": 7.36198347107438e-07,
+      "logits/chosen": -0.10009765625,
+      "logits/rejected": -0.21875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3724,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 4.875,
+      "rewards/rejected": -3.515625,
+      "step": 218
+    },
+    {
+      "epoch": 0.2687116564417178,
+      "grad_norm": 36.99776737120032,
+      "learning_rate": 7.358677685950413e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3041,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.4140625,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -3.875,
+      "step": 219
+    },
+    {
+      "epoch": 0.26993865030674846,
+      "grad_norm": 49.4734461720144,
+      "learning_rate": 7.355371900826446e-07,
+      "logits/chosen": -0.05078125,
+      "logits/rejected": -0.1318359375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3781,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 4.375,
+      "rewards/rejected": -3.421875,
+      "step": 220
+    },
+    {
+      "epoch": 0.27116564417177913,
+      "grad_norm": 50.14968062203286,
+      "learning_rate": 7.352066115702479e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -292.0,
+      "loss": 0.3828,
+      "rewards/accuracies": 0.765625,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 4.9375,
+      "rewards/rejected": -3.796875,
+      "step": 221
+    },
+    {
+      "epoch": 0.2723926380368098,
+      "grad_norm": 35.95574239692987,
+      "learning_rate": 7.348760330578512e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2649,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.09375,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.15625,
+      "step": 222
+    },
+    {
+      "epoch": 0.27361963190184047,
+      "grad_norm": 40.443295860960646,
+      "learning_rate": 7.345454545454545e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.314453125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2888,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.75,
+      "step": 223
+    },
+    {
+      "epoch": 0.2748466257668712,
+      "grad_norm": 42.19474227390439,
+      "learning_rate": 7.342148760330579e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -446.0,
+      "logps/rejected": -334.0,
+      "loss": 0.288,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.09375,
+      "step": 224
+    },
+    {
+      "epoch": 0.27607361963190186,
+      "grad_norm": 37.9965018970918,
+      "learning_rate": 7.338842975206612e-07,
+      "logits/chosen": -0.10546875,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -336.0,
+      "logps/rejected": -304.0,
+      "loss": 0.261,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.40625,
+      "step": 225
+    },
+    {
+      "epoch": 0.2773006134969325,
+      "grad_norm": 38.80764018954137,
+      "learning_rate": 7.335537190082644e-07,
+      "logits/chosen": -0.119140625,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3311,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -3.890625,
+      "step": 226
+    },
+    {
+      "epoch": 0.2785276073619632,
+      "grad_norm": 42.55177823975192,
+      "learning_rate": 7.332231404958677e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3198,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.40625,
+      "step": 227
+    },
+    {
+      "epoch": 0.27975460122699386,
+      "grad_norm": 47.83061041994659,
+      "learning_rate": 7.32892561983471e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3258,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -4.125,
+      "step": 228
+    },
+    {
+      "epoch": 0.2809815950920245,
+      "grad_norm": 36.471929966124165,
+      "learning_rate": 7.325619834710743e-07,
+      "logits/chosen": -0.07373046875,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2558,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.59375,
+      "step": 229
+    },
+    {
+      "epoch": 0.2822085889570552,
+      "grad_norm": 36.373451110997884,
+      "learning_rate": 7.322314049586777e-07,
+      "logits/chosen": -0.06884765625,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2454,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.28125,
+      "step": 230
+    },
+    {
+      "epoch": 0.28343558282208586,
+      "grad_norm": 44.837599338102365,
+      "learning_rate": 7.31900826446281e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3895,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -3.625,
+      "step": 231
+    },
+    {
+      "epoch": 0.2846625766871166,
+      "grad_norm": 40.26904672106654,
+      "learning_rate": 7.315702479338843e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3097,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -4.0625,
+      "step": 232
+    },
+    {
+      "epoch": 0.28588957055214725,
+      "grad_norm": 37.854698103189165,
+      "learning_rate": 7.312396694214876e-07,
+      "logits/chosen": -0.10888671875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -288.0,
+      "loss": 0.3055,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.0625,
+      "step": 233
+    },
+    {
+      "epoch": 0.2871165644171779,
+      "grad_norm": 38.018318477115464,
+      "learning_rate": 7.309090909090909e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2977,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -3.71875,
+      "step": 234
+    },
+    {
+      "epoch": 0.2883435582822086,
+      "grad_norm": 37.47355891059506,
+      "learning_rate": 7.305785123966941e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -320.0,
+      "loss": 0.284,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.0625,
+      "step": 235
+    },
+    {
+      "epoch": 0.28957055214723926,
+      "grad_norm": 53.75756263750908,
+      "learning_rate": 7.302479338842975e-07,
+      "logits/chosen": -0.0238037109375,
+      "logits/rejected": -0.2080078125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -368.0,
+      "loss": 0.4235,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.1875,
+      "step": 236
+    },
+    {
+      "epoch": 0.2907975460122699,
+      "grad_norm": 34.35888091553625,
+      "learning_rate": 7.299173553719008e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2606,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.09375,
+      "step": 237
+    },
+    {
+      "epoch": 0.2920245398773006,
+      "grad_norm": 42.88801905010467,
+      "learning_rate": 7.295867768595041e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3538,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.7578125,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -3.90625,
+      "step": 238
+    },
+    {
+      "epoch": 0.29325153374233126,
+      "grad_norm": 42.29179859412283,
+      "learning_rate": 7.292561983471074e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3291,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.15625,
+      "step": 239
+    },
+    {
+      "epoch": 0.294478527607362,
+      "grad_norm": 37.481940316076816,
+      "learning_rate": 7.289256198347107e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -360.0,
+      "loss": 0.2328,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.890625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -4.6875,
+      "step": 240
+    },
+    {
+      "epoch": 0.29570552147239265,
+      "grad_norm": 37.13709897048156,
+      "learning_rate": 7.28595041322314e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2698,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.25,
+      "step": 241
+    },
+    {
+      "epoch": 0.2969325153374233,
+      "grad_norm": 30.985673644793227,
+      "learning_rate": 7.282644628099174e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -318.0,
+      "loss": 0.228,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.84375,
+      "step": 242
+    },
+    {
+      "epoch": 0.298159509202454,
+      "grad_norm": 53.9279211325065,
+      "learning_rate": 7.279338842975206e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -364.0,
+      "loss": 0.3208,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.34375,
+      "step": 243
+    },
+    {
+      "epoch": 0.29938650306748466,
+      "grad_norm": 44.157193824232536,
+      "learning_rate": 7.276033057851239e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -346.0,
+      "loss": 0.4473,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.125,
+      "step": 244
+    },
+    {
+      "epoch": 0.3006134969325153,
+      "grad_norm": 36.965116220324916,
+      "learning_rate": 7.272727272727272e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2506,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.46875,
+      "step": 245
+    },
+    {
+      "epoch": 0.301840490797546,
+      "grad_norm": 35.79905789258026,
+      "learning_rate": 7.269421487603305e-07,
+      "logits/chosen": -0.2158203125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.292,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.078125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.40625,
+      "step": 246
+    },
+    {
+      "epoch": 0.3030674846625767,
+      "grad_norm": 46.030369589419166,
+      "learning_rate": 7.266115702479338e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3237,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.75,
+      "step": 247
+    },
+    {
+      "epoch": 0.3042944785276074,
+      "grad_norm": 38.420936945371935,
+      "learning_rate": 7.262809917355372e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2497,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.28125,
+      "step": 248
+    },
+    {
+      "epoch": 0.30552147239263805,
+      "grad_norm": 42.67532255005307,
+      "learning_rate": 7.259504132231405e-07,
+      "logits/chosen": -0.25390625,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -476.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2862,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.328125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.3125,
+      "step": 249
+    },
+    {
+      "epoch": 0.3067484662576687,
+      "grad_norm": 40.02525959257945,
+      "learning_rate": 7.256198347107438e-07,
+      "logits/chosen": -0.34375,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -342.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3299,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -4.40625,
+      "step": 250
+    },
+    {
+      "epoch": 0.3079754601226994,
+      "grad_norm": 43.266430285019,
+      "learning_rate": 7.252892561983471e-07,
+      "logits/chosen": -0.041015625,
+      "logits/rejected": -0.2138671875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3567,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.75,
+      "step": 251
+    },
+    {
+      "epoch": 0.30920245398773005,
+      "grad_norm": 37.89906733852128,
+      "learning_rate": 7.249586776859503e-07,
+      "logits/chosen": -0.205078125,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2982,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.7734375,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -3.859375,
+      "step": 252
+    },
+    {
+      "epoch": 0.3104294478527607,
+      "grad_norm": 33.28018552295872,
+      "learning_rate": 7.246280991735536e-07,
+      "logits/chosen": -0.208984375,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2522,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.46875,
+      "step": 253
+    },
+    {
+      "epoch": 0.3116564417177914,
+      "grad_norm": 42.254563825083,
+      "learning_rate": 7.24297520661157e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3681,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.15625,
+      "step": 254
+    },
+    {
+      "epoch": 0.3128834355828221,
+      "grad_norm": 39.895653943663504,
+      "learning_rate": 7.239669421487603e-07,
+      "logits/chosen": -0.1396484375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3481,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.8359375,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.46875,
+      "step": 255
+    },
+    {
+      "epoch": 0.3141104294478528,
+      "grad_norm": 37.11298367671905,
+      "learning_rate": 7.236363636363636e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -348.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2374,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.34375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.0625,
+      "step": 256
+    },
+    {
+      "epoch": 0.31533742331288345,
+      "grad_norm": 44.74542880830349,
+      "learning_rate": 7.233057851239669e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -426.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3118,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -3.546875,
+      "step": 257
+    },
+    {
+      "epoch": 0.3165644171779141,
+      "grad_norm": 34.05504946265699,
+      "learning_rate": 7.229752066115702e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2467,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.8984375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -3.859375,
+      "step": 258
+    },
+    {
+      "epoch": 0.3177914110429448,
+      "grad_norm": 35.06646898219457,
+      "learning_rate": 7.226446280991736e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -344.0,
+      "loss": 0.278,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.96875,
+      "step": 259
+    },
+    {
+      "epoch": 0.31901840490797545,
+      "grad_norm": 48.15030024240453,
+      "learning_rate": 7.223140495867769e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3099,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -3.640625,
+      "step": 260
+    },
+    {
+      "epoch": 0.3202453987730061,
+      "grad_norm": 43.183782239877885,
+      "learning_rate": 7.219834710743801e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3207,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.71875,
+      "step": 261
+    },
+    {
+      "epoch": 0.3214723926380368,
+      "grad_norm": 43.81473087786423,
+      "learning_rate": 7.216528925619834e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3077,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -3.484375,
+      "step": 262
+    },
+    {
+      "epoch": 0.3226993865030675,
+      "grad_norm": 45.763039013247074,
+      "learning_rate": 7.213223140495867e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -304.0,
+      "loss": 0.345,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 4.5,
+      "rewards/rejected": -3.265625,
+      "step": 263
+    },
+    {
+      "epoch": 0.3239263803680982,
+      "grad_norm": 41.54807658858699,
+      "learning_rate": 7.2099173553719e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -288.0,
+      "loss": 0.3193,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -3.234375,
+      "step": 264
+    },
+    {
+      "epoch": 0.32515337423312884,
+      "grad_norm": 43.72770395576322,
+      "learning_rate": 7.206611570247934e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -266.0,
+      "loss": 0.296,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 4.96875,
+      "rewards/rejected": -3.515625,
+      "step": 265
+    },
+    {
+      "epoch": 0.3263803680981595,
+      "grad_norm": 49.68472762312363,
+      "learning_rate": 7.203305785123967e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -316.0,
+      "logps/rejected": -292.0,
+      "loss": 0.41,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 0.68359375,
+      "rewards/margins": 3.875,
+      "rewards/rejected": -3.1875,
+      "step": 266
+    },
+    {
+      "epoch": 0.3276073619631902,
+      "grad_norm": 54.21686609921306,
+      "learning_rate": 7.2e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -296.0,
+      "loss": 0.3765,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.828125,
+      "rewards/margins": 4.8125,
+      "rewards/rejected": -2.984375,
+      "step": 267
+    },
+    {
+      "epoch": 0.32883435582822085,
+      "grad_norm": 36.013473849533085,
+      "learning_rate": 7.196694214876033e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2353,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.328125,
+      "step": 268
+    },
+    {
+      "epoch": 0.3300613496932515,
+      "grad_norm": 37.07985683341205,
+      "learning_rate": 7.193388429752066e-07,
+      "logits/chosen": -0.193359375,
+      "logits/rejected": -0.5,
+      "logps/chosen": -400.0,
+      "logps/rejected": -292.0,
+      "loss": 0.229,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -3.46875,
+      "step": 269
+    },
+    {
+      "epoch": 0.3312883435582822,
+      "grad_norm": 43.39229519288019,
+      "learning_rate": 7.190082644628098e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -326.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3693,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 4.1875,
+      "rewards/rejected": -2.875,
+      "step": 270
+    },
+    {
+      "epoch": 0.3325153374233129,
+      "grad_norm": 33.45853348793219,
+      "learning_rate": 7.186776859504132e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2511,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.34375,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.234375,
+      "step": 271
+    },
+    {
+      "epoch": 0.3337423312883436,
+      "grad_norm": 36.86544969536763,
+      "learning_rate": 7.183471074380165e-07,
+      "logits/chosen": -0.1025390625,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2449,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.640625,
+      "step": 272
+    },
+    {
+      "epoch": 0.33496932515337424,
+      "grad_norm": 45.60464590137911,
+      "learning_rate": 7.180165289256198e-07,
+      "logits/chosen": -0.05029296875,
+      "logits/rejected": -0.1357421875,
+      "logps/chosen": -334.0,
+      "logps/rejected": -292.0,
+      "loss": 0.4336,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 4.59375,
+      "rewards/rejected": -3.4375,
+      "step": 273
+    },
+    {
+      "epoch": 0.3361963190184049,
+      "grad_norm": 38.72003382453235,
+      "learning_rate": 7.176859504132231e-07,
+      "logits/chosen": -0.06591796875,
+      "logits/rejected": -0.21875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3223,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.359375,
+      "step": 274
+    },
+    {
+      "epoch": 0.3374233128834356,
+      "grad_norm": 41.46753745108232,
+      "learning_rate": 7.173553719008264e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -316.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3348,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -3.5,
+      "step": 275
+    },
+    {
+      "epoch": 0.33865030674846625,
+      "grad_norm": 29.942801211334718,
+      "learning_rate": 7.170247933884297e-07,
+      "logits/chosen": -0.056396484375,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2861,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -3.84375,
+      "step": 276
+    },
+    {
+      "epoch": 0.3398773006134969,
+      "grad_norm": 36.649252691867474,
+      "learning_rate": 7.166942148760331e-07,
+      "logits/chosen": -0.2265625,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2238,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.3125,
+      "step": 277
+    },
+    {
+      "epoch": 0.3411042944785276,
+      "grad_norm": 41.64219712301306,
+      "learning_rate": 7.163636363636364e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2881,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.125,
+      "step": 278
+    },
+    {
+      "epoch": 0.3423312883435583,
+      "grad_norm": 25.028100153022006,
+      "learning_rate": 7.160330578512396e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1819,
+      "rewards/accuracies": 0.9453125,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.875,
+      "step": 279
+    },
+    {
+      "epoch": 0.34355828220858897,
+      "grad_norm": 43.34524582256192,
+      "learning_rate": 7.157024793388429e-07,
+      "logits/chosen": -0.177734375,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3811,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.625,
+      "step": 280
+    },
+    {
+      "epoch": 0.34478527607361964,
+      "grad_norm": 39.57499610045821,
+      "learning_rate": 7.153719008264462e-07,
+      "logits/chosen": -0.12158203125,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3441,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -3.9375,
+      "step": 281
+    },
+    {
+      "epoch": 0.3460122699386503,
+      "grad_norm": 32.558914994483004,
+      "learning_rate": 7.150413223140495e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -340.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2162,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.4765625,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.0625,
+      "step": 282
+    },
+    {
+      "epoch": 0.347239263803681,
+      "grad_norm": 33.09320412279389,
+      "learning_rate": 7.147107438016529e-07,
+      "logits/chosen": -0.314453125,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -350.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2537,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.984375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.5,
+      "step": 283
+    },
+    {
+      "epoch": 0.34846625766871164,
+      "grad_norm": 40.734540729970675,
+      "learning_rate": 7.143801652892562e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -310.0,
+      "loss": 0.272,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.84375,
+      "step": 284
+    },
+    {
+      "epoch": 0.3496932515337423,
+      "grad_norm": 30.9410258264426,
+      "learning_rate": 7.140495867768595e-07,
+      "logits/chosen": -0.21484375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2413,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.96875,
+      "step": 285
+    },
+    {
+      "epoch": 0.350920245398773,
+      "grad_norm": 37.246332837421065,
+      "learning_rate": 7.137190082644628e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3337,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.09375,
+      "step": 286
+    },
+    {
+      "epoch": 0.3521472392638037,
+      "grad_norm": 40.87355956326826,
+      "learning_rate": 7.133884297520661e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2981,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.68359375,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.9375,
+      "step": 287
+    },
+    {
+      "epoch": 0.35337423312883437,
+      "grad_norm": 37.148615900618886,
+      "learning_rate": 7.130578512396693e-07,
+      "logits/chosen": -0.05908203125,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3085,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -4.59375,
+      "step": 288
+    },
+    {
+      "epoch": 0.35460122699386504,
+      "grad_norm": 29.32351813266847,
+      "learning_rate": 7.127272727272727e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -316.0,
+      "loss": 0.1987,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.34375,
+      "step": 289
+    },
+    {
+      "epoch": 0.3558282208588957,
+      "grad_norm": 34.03478605235303,
+      "learning_rate": 7.12396694214876e-07,
+      "logits/chosen": -0.359375,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2024,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.71875,
+      "step": 290
+    },
+    {
+      "epoch": 0.3570552147239264,
+      "grad_norm": 38.57928856233812,
+      "learning_rate": 7.120661157024793e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3224,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -4.5,
+      "step": 291
+    },
+    {
+      "epoch": 0.35828220858895704,
+      "grad_norm": 36.5133266014805,
+      "learning_rate": 7.117355371900826e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.484375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2451,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.03125,
+      "step": 292
+    },
+    {
+      "epoch": 0.3595092024539877,
+      "grad_norm": 31.073871401773395,
+      "learning_rate": 7.114049586776859e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2162,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.25,
+      "step": 293
+    },
+    {
+      "epoch": 0.36073619631901843,
+      "grad_norm": 31.419020907335653,
+      "learning_rate": 7.110743801652892e-07,
+      "logits/chosen": -0.427734375,
+      "logits/rejected": -0.578125,
+      "logps/chosen": -464.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2389,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.3828125,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.53125,
+      "step": 294
+    },
+    {
+      "epoch": 0.3619631901840491,
+      "grad_norm": 33.66575248055817,
+      "learning_rate": 7.107438016528926e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2242,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.93359375,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.03125,
+      "step": 295
+    },
+    {
+      "epoch": 0.36319018404907977,
+      "grad_norm": 40.06405186289625,
+      "learning_rate": 7.104132231404959e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3209,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.1875,
+      "step": 296
+    },
+    {
+      "epoch": 0.36441717791411044,
+      "grad_norm": 30.371449913820346,
+      "learning_rate": 7.100826446280991e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -352.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2354,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.5,
+      "step": 297
+    },
+    {
+      "epoch": 0.3656441717791411,
+      "grad_norm": 40.780831154790015,
+      "learning_rate": 7.097520661157024e-07,
+      "logits/chosen": -0.126953125,
+      "logits/rejected": -0.23046875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2901,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 5.25,
+      "rewards/rejected": -3.890625,
+      "step": 298
+    },
+    {
+      "epoch": 0.36687116564417177,
+      "grad_norm": 48.39132901882327,
+      "learning_rate": 7.094214876033057e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3427,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -4.0625,
+      "step": 299
+    },
+    {
+      "epoch": 0.36809815950920244,
+      "grad_norm": 39.25730980280545,
+      "learning_rate": 7.09090909090909e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3287,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.21875,
+      "step": 300
+    },
+    {
+      "epoch": 0.3693251533742331,
+      "grad_norm": 34.87785766178715,
+      "learning_rate": 7.087603305785124e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -448.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2823,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.1875,
+      "step": 301
+    },
+    {
+      "epoch": 0.37055214723926383,
+      "grad_norm": 30.464291800676158,
+      "learning_rate": 7.084297520661157e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2182,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -4.15625,
+      "step": 302
+    },
+    {
+      "epoch": 0.3717791411042945,
+      "grad_norm": 45.78814856158198,
+      "learning_rate": 7.08099173553719e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3512,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.40625,
+      "step": 303
+    },
+    {
+      "epoch": 0.37300613496932516,
+      "grad_norm": 33.26834608933062,
+      "learning_rate": 7.077685950413223e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2448,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.53125,
+      "step": 304
+    },
+    {
+      "epoch": 0.37423312883435583,
+      "grad_norm": 35.609604648743236,
+      "learning_rate": 7.074380165289256e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2744,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.625,
+      "step": 305
+    },
+    {
+      "epoch": 0.3754601226993865,
+      "grad_norm": 43.837358024220855,
+      "learning_rate": 7.071074380165288e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2737,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.5,
+      "step": 306
+    },
+    {
+      "epoch": 0.37668711656441717,
+      "grad_norm": 39.17650253748135,
+      "learning_rate": 7.067768595041322e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3713,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.71875,
+      "step": 307
+    },
+    {
+      "epoch": 0.37791411042944784,
+      "grad_norm": 36.807676017880326,
+      "learning_rate": 7.064462809917355e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3165,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.94140625,
+      "rewards/margins": 5.0,
+      "rewards/rejected": -4.0625,
+      "step": 308
+    },
+    {
+      "epoch": 0.3791411042944785,
+      "grad_norm": 32.288551419968854,
+      "learning_rate": 7.061157024793388e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2283,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.6484375,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -4.90625,
+      "step": 309
+    },
+    {
+      "epoch": 0.3803680981595092,
+      "grad_norm": 47.130716517812814,
+      "learning_rate": 7.057851239669421e-07,
+      "logits/chosen": -0.0908203125,
+      "logits/rejected": -0.1884765625,
+      "logps/chosen": -312.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3596,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.15234375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.6875,
+      "step": 310
+    },
+    {
+      "epoch": 0.3815950920245399,
+      "grad_norm": 34.12148114082454,
+      "learning_rate": 7.054545454545454e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.4453125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2543,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -5.375,
+      "step": 311
+    },
+    {
+      "epoch": 0.38282208588957056,
+      "grad_norm": 44.35564676375722,
+      "learning_rate": 7.051239669421488e-07,
+      "logits/chosen": -0.125,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -362.0,
+      "loss": 0.4044,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -4.5,
+      "step": 312
+    },
+    {
+      "epoch": 0.38404907975460123,
+      "grad_norm": 33.82923751151301,
+      "learning_rate": 7.047933884297521e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.412109375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2685,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.8828125,
+      "rewards/margins": 7.28125,
+      "rewards/rejected": -5.375,
+      "step": 313
+    },
+    {
+      "epoch": 0.3852760736196319,
+      "grad_norm": 33.17604766084792,
+      "learning_rate": 7.044628099173554e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -456.0,
+      "logps/rejected": -344.0,
+      "loss": 0.1847,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 2.25,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -5.71875,
+      "step": 314
+    },
+    {
+      "epoch": 0.38650306748466257,
+      "grad_norm": 38.90508166705689,
+      "learning_rate": 7.041322314049586e-07,
+      "logits/chosen": -0.06396484375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -286.0,
+      "loss": 0.2412,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.5546875,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.21875,
+      "step": 315
+    },
+    {
+      "epoch": 0.38773006134969323,
+      "grad_norm": 31.950078808944912,
+      "learning_rate": 7.038016528925619e-07,
+      "logits/chosen": -0.06982421875,
+      "logits/rejected": -0.2158203125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2551,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.4375,
+      "step": 316
+    },
+    {
+      "epoch": 0.3889570552147239,
+      "grad_norm": 34.528377745666106,
+      "learning_rate": 7.034710743801652e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2296,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.875,
+      "step": 317
+    },
+    {
+      "epoch": 0.3901840490797546,
+      "grad_norm": 46.9506582652498,
+      "learning_rate": 7.031404958677686e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3503,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.71875,
+      "step": 318
+    },
+    {
+      "epoch": 0.3914110429447853,
+      "grad_norm": 26.867054074267944,
+      "learning_rate": 7.028099173553719e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1701,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 7.4375,
+      "rewards/rejected": -6.125,
+      "step": 319
+    },
+    {
+      "epoch": 0.39263803680981596,
+      "grad_norm": 40.540053142562215,
+      "learning_rate": 7.024793388429752e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -318.0,
+      "logps/rejected": -274.0,
+      "loss": 0.3046,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -5.3125,
+      "step": 320
+    },
+    {
+      "epoch": 0.39386503067484663,
+      "grad_norm": 46.49642012587718,
+      "learning_rate": 7.021487603305785e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -384.0,
+      "loss": 0.2961,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.96875,
+      "step": 321
+    },
+    {
+      "epoch": 0.3950920245398773,
+      "grad_norm": 33.72622281793196,
+      "learning_rate": 7.018181818181818e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2496,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.3125,
+      "step": 322
+    },
+    {
+      "epoch": 0.39631901840490796,
+      "grad_norm": 39.04984997032008,
+      "learning_rate": 7.014876033057851e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -350.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3837,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -4.3125,
+      "step": 323
+    },
+    {
+      "epoch": 0.39754601226993863,
+      "grad_norm": 33.57036215026917,
+      "learning_rate": 7.011570247933884e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -352.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2714,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.5,
+      "step": 324
+    },
+    {
+      "epoch": 0.3987730061349693,
+      "grad_norm": 45.64995516248102,
+      "learning_rate": 7.008264462809917e-07,
+      "logits/chosen": -0.138671875,
+      "logits/rejected": -0.2177734375,
+      "logps/chosen": -328.0,
+      "logps/rejected": -290.0,
+      "loss": 0.4112,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.6875,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -4.53125,
+      "step": 325
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 34.28943945713478,
+      "learning_rate": 7.00495867768595e-07,
+      "logits/chosen": -0.310546875,
+      "logits/rejected": -0.462890625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2004,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.09375,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -5.03125,
+      "step": 326
+    },
+    {
+      "epoch": 0.4012269938650307,
+      "grad_norm": 48.50197304580951,
+      "learning_rate": 7.001652892561983e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3949,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.21875,
+      "rewards/rejected": -3.6875,
+      "step": 327
+    },
+    {
+      "epoch": 0.40245398773006136,
+      "grad_norm": 40.72777385232644,
+      "learning_rate": 6.998347107438016e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -272.0,
+      "loss": 0.3146,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -4.3125,
+      "step": 328
+    },
+    {
+      "epoch": 0.403680981595092,
+      "grad_norm": 35.55525281765151,
+      "learning_rate": 6.995041322314049e-07,
+      "logits/chosen": -0.1787109375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -454.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2936,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -4.90625,
+      "step": 329
+    },
+    {
+      "epoch": 0.4049079754601227,
+      "grad_norm": 42.0221302571432,
+      "learning_rate": 6.991735537190083e-07,
+      "logits/chosen": -0.3125,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2435,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9765625,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -4.71875,
+      "step": 330
+    },
+    {
+      "epoch": 0.40613496932515336,
+      "grad_norm": 32.84700653520424,
+      "learning_rate": 6.988429752066116e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2501,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.9375,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -4.375,
+      "step": 331
+    },
+    {
+      "epoch": 0.40736196319018403,
+      "grad_norm": 27.379927529722597,
+      "learning_rate": 6.985123966942148e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2017,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.46875,
+      "step": 332
+    },
+    {
+      "epoch": 0.4085889570552147,
+      "grad_norm": 36.70659629756495,
+      "learning_rate": 6.981818181818181e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3078,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -3.84375,
+      "step": 333
+    },
+    {
+      "epoch": 0.4098159509202454,
+      "grad_norm": 34.15971280742034,
+      "learning_rate": 6.978512396694214e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -338.0,
+      "loss": 0.273,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -4.5,
+      "step": 334
+    },
+    {
+      "epoch": 0.4110429447852761,
+      "grad_norm": 37.2375068536928,
+      "learning_rate": 6.975206611570247e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2248,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.03125,
+      "step": 335
+    },
+    {
+      "epoch": 0.41226993865030676,
+      "grad_norm": 36.84981287485191,
+      "learning_rate": 6.971900826446281e-07,
+      "logits/chosen": -0.294921875,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2764,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.21875,
+      "step": 336
+    },
+    {
+      "epoch": 0.4134969325153374,
+      "grad_norm": 46.94122451213228,
+      "learning_rate": 6.968595041322314e-07,
+      "logits/chosen": -0.09375,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -344.0,
+      "logps/rejected": -322.0,
+      "loss": 0.389,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.921875,
+      "step": 337
+    },
+    {
+      "epoch": 0.4147239263803681,
+      "grad_norm": 38.247859512270246,
+      "learning_rate": 6.965289256198347e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3295,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.0,
+      "step": 338
+    },
+    {
+      "epoch": 0.41595092024539876,
+      "grad_norm": 40.84741702132196,
+      "learning_rate": 6.96198347107438e-07,
+      "logits/chosen": -0.07470703125,
+      "logits/rejected": -0.2109375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3275,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.15625,
+      "step": 339
+    },
+    {
+      "epoch": 0.4171779141104294,
+      "grad_norm": 43.700613878566976,
+      "learning_rate": 6.958677685950413e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2879,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.09375,
+      "step": 340
+    },
+    {
+      "epoch": 0.41840490797546015,
+      "grad_norm": 31.54531879054049,
+      "learning_rate": 6.955371900826445e-07,
+      "logits/chosen": -0.21875,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2805,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.65625,
+      "step": 341
+    },
+    {
+      "epoch": 0.4196319018404908,
+      "grad_norm": 39.0902727508844,
+      "learning_rate": 6.952066115702479e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -328.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3322,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.1875,
+      "step": 342
+    },
+    {
+      "epoch": 0.4208588957055215,
+      "grad_norm": 39.32051666516945,
+      "learning_rate": 6.948760330578512e-07,
+      "logits/chosen": -0.1962890625,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3545,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -3.921875,
+      "step": 343
+    },
+    {
+      "epoch": 0.42208588957055215,
+      "grad_norm": 42.12526028461151,
+      "learning_rate": 6.945454545454545e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3584,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.0625,
+      "step": 344
+    },
+    {
+      "epoch": 0.4233128834355828,
+      "grad_norm": 39.02371998906191,
+      "learning_rate": 6.942148760330578e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3288,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.3125,
+      "step": 345
+    },
+    {
+      "epoch": 0.4245398773006135,
+      "grad_norm": 41.19695447421548,
+      "learning_rate": 6.938842975206611e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -332.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3569,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 5.09375,
+      "rewards/rejected": -4.125,
+      "step": 346
+    },
+    {
+      "epoch": 0.42576687116564416,
+      "grad_norm": 28.472619739287644,
+      "learning_rate": 6.935537190082645e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -360.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2332,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.25,
+      "step": 347
+    },
+    {
+      "epoch": 0.4269938650306748,
+      "grad_norm": 33.19419614917591,
+      "learning_rate": 6.932231404958678e-07,
+      "logits/chosen": -0.24609375,
+      "logits/rejected": -0.455078125,
+      "logps/chosen": -410.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2594,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.91796875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.59375,
+      "step": 348
+    },
+    {
+      "epoch": 0.42822085889570555,
+      "grad_norm": 38.111059722172605,
+      "learning_rate": 6.928925619834711e-07,
+      "logits/chosen": -0.2412109375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3255,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.34375,
+      "step": 349
+    },
+    {
+      "epoch": 0.4294478527607362,
+      "grad_norm": 37.49784569992658,
+      "learning_rate": 6.925619834710743e-07,
+      "logits/chosen": -0.0986328125,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2857,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.28125,
+      "step": 350
+    },
+    {
+      "epoch": 0.4306748466257669,
+      "grad_norm": 42.647163206760126,
+      "learning_rate": 6.922314049586776e-07,
+      "logits/chosen": -0.259765625,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -336.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3505,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.5078125,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -4.78125,
+      "step": 351
+    },
+    {
+      "epoch": 0.43190184049079755,
+      "grad_norm": 36.26120743659289,
+      "learning_rate": 6.919008264462809e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3453,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.03125,
+      "step": 352
+    },
+    {
+      "epoch": 0.4331288343558282,
+      "grad_norm": 30.933441803520967,
+      "learning_rate": 6.915702479338843e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.3984375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2406,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.484375,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.96875,
+      "step": 353
+    },
+    {
+      "epoch": 0.4343558282208589,
+      "grad_norm": 34.874499707263155,
+      "learning_rate": 6.912396694214876e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2963,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.03125,
+      "step": 354
+    },
+    {
+      "epoch": 0.43558282208588955,
+      "grad_norm": 39.848840136754845,
+      "learning_rate": 6.909090909090909e-07,
+      "logits/chosen": -0.296875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -340.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3581,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -4.53125,
+      "step": 355
+    },
+    {
+      "epoch": 0.4368098159509202,
+      "grad_norm": 44.03875893573232,
+      "learning_rate": 6.905785123966942e-07,
+      "logits/chosen": -0.0966796875,
+      "logits/rejected": -0.240234375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3406,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.03125,
+      "step": 356
+    },
+    {
+      "epoch": 0.43803680981595094,
+      "grad_norm": 36.68968799262988,
+      "learning_rate": 6.902479338842975e-07,
+      "logits/chosen": -0.3515625,
+      "logits/rejected": -0.50390625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2926,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -4.9375,
+      "step": 357
+    },
+    {
+      "epoch": 0.4392638036809816,
+      "grad_norm": 36.204158508103596,
+      "learning_rate": 6.899173553719008e-07,
+      "logits/chosen": -0.30078125,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2673,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.40625,
+      "step": 358
+    },
+    {
+      "epoch": 0.4404907975460123,
+      "grad_norm": 40.16138545949422,
+      "learning_rate": 6.89586776859504e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.953125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.09375,
+      "step": 359
+    },
+    {
+      "epoch": 0.44171779141104295,
+      "grad_norm": 35.21609251200673,
+      "learning_rate": 6.892561983471074e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2614,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.9375,
+      "step": 360
+    },
+    {
+      "epoch": 0.4429447852760736,
+      "grad_norm": 33.95190421090596,
+      "learning_rate": 6.889256198347107e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -426.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2599,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -4.8125,
+      "step": 361
+    },
+    {
+      "epoch": 0.4441717791411043,
+      "grad_norm": 32.41395226001234,
+      "learning_rate": 6.88595041322314e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2625,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -4.75,
+      "step": 362
+    },
+    {
+      "epoch": 0.44539877300613495,
+      "grad_norm": 31.06408733284358,
+      "learning_rate": 6.882644628099173e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2362,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.375,
+      "step": 363
+    },
+    {
+      "epoch": 0.4466257668711656,
+      "grad_norm": 40.778000027843554,
+      "learning_rate": 6.879338842975206e-07,
+      "logits/chosen": -0.07470703125,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2885,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.28125,
+      "step": 364
+    },
+    {
+      "epoch": 0.44785276073619634,
+      "grad_norm": 30.8545654380382,
+      "learning_rate": 6.87603305785124e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -298.0,
+      "loss": 0.2419,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.53125,
+      "step": 365
+    },
+    {
+      "epoch": 0.449079754601227,
+      "grad_norm": 35.25923514975111,
+      "learning_rate": 6.872727272727273e-07,
+      "logits/chosen": -0.11376953125,
+      "logits/rejected": -0.21484375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3324,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.71875,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -3.703125,
+      "step": 366
+    },
+    {
+      "epoch": 0.4503067484662577,
+      "grad_norm": 29.75727838549342,
+      "learning_rate": 6.869421487603306e-07,
+      "logits/chosen": -0.2490234375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2265,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.25,
+      "step": 367
+    },
+    {
+      "epoch": 0.45153374233128835,
+      "grad_norm": 41.34860966923763,
+      "learning_rate": 6.866115702479338e-07,
+      "logits/chosen": -0.060546875,
+      "logits/rejected": -0.11328125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3251,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.1875,
+      "step": 368
+    },
+    {
+      "epoch": 0.452760736196319,
+      "grad_norm": 39.802701196225414,
+      "learning_rate": 6.862809917355371e-07,
+      "logits/chosen": -0.08349609375,
+      "logits/rejected": -0.19140625,
+      "logps/chosen": -340.0,
+      "logps/rejected": -286.0,
+      "loss": 0.3103,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 5.03125,
+      "rewards/rejected": -4.375,
+      "step": 369
+    },
+    {
+      "epoch": 0.4539877300613497,
+      "grad_norm": 35.44626532768664,
+      "learning_rate": 6.859504132231404e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3001,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.0,
+      "step": 370
+    },
+    {
+      "epoch": 0.45521472392638035,
+      "grad_norm": 34.081640470685045,
+      "learning_rate": 6.856198347107438e-07,
+      "logits/chosen": -0.2314453125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -458.0,
+      "logps/rejected": -356.0,
+      "loss": 0.268,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.734375,
+      "step": 371
+    },
+    {
+      "epoch": 0.456441717791411,
+      "grad_norm": 32.10146522336104,
+      "learning_rate": 6.852892561983471e-07,
+      "logits/chosen": -0.14453125,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2227,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.15625,
+      "step": 372
+    },
+    {
+      "epoch": 0.45766871165644174,
+      "grad_norm": 38.40212256087746,
+      "learning_rate": 6.849586776859504e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -360.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3491,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.88671875,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -4.40625,
+      "step": 373
+    },
+    {
+      "epoch": 0.4588957055214724,
+      "grad_norm": 33.89078532536025,
+      "learning_rate": 6.846280991735537e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -310.0,
+      "loss": 0.2439,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 2.03125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.25,
+      "step": 374
+    },
+    {
+      "epoch": 0.4601226993865031,
+      "grad_norm": 38.36603856770271,
+      "learning_rate": 6.84297520661157e-07,
+      "logits/chosen": -0.24609375,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2743,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -3.78125,
+      "step": 375
+    },
+    {
+      "epoch": 0.46134969325153374,
+      "grad_norm": 32.73897251445023,
+      "learning_rate": 6.839669421487603e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -342.0,
+      "loss": 0.231,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -3.9375,
+      "step": 376
+    },
+    {
+      "epoch": 0.4625766871165644,
+      "grad_norm": 36.44881761827906,
+      "learning_rate": 6.836363636363636e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -350.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2817,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.8125,
+      "step": 377
+    },
+    {
+      "epoch": 0.4638036809815951,
+      "grad_norm": 36.92866834323212,
+      "learning_rate": 6.833057851239669e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3001,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.5,
+      "step": 378
+    },
+    {
+      "epoch": 0.46503067484662575,
+      "grad_norm": 32.73144242036404,
+      "learning_rate": 6.829752066115702e-07,
+      "logits/chosen": -0.27734375,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -408.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2471,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.7578125,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.0625,
+      "step": 379
+    },
+    {
+      "epoch": 0.4662576687116564,
+      "grad_norm": 34.6763093414022,
+      "learning_rate": 6.826446280991735e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2659,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.5546875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -4.96875,
+      "step": 380
+    },
+    {
+      "epoch": 0.46748466257668714,
+      "grad_norm": 34.98446202126656,
+      "learning_rate": 6.823140495867768e-07,
+      "logits/chosen": -0.2734375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3193,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.59375,
+      "step": 381
+    },
+    {
+      "epoch": 0.4687116564417178,
+      "grad_norm": 33.068661800772716,
+      "learning_rate": 6.819834710743801e-07,
+      "logits/chosen": -0.279296875,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.40625,
+      "step": 382
+    },
+    {
+      "epoch": 0.4699386503067485,
+      "grad_norm": 37.187464031340724,
+      "learning_rate": 6.816528925619835e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -378.0,
+      "loss": 0.33,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.625,
+      "step": 383
+    },
+    {
+      "epoch": 0.47116564417177914,
+      "grad_norm": 35.133860722289874,
+      "learning_rate": 6.813223140495868e-07,
+      "logits/chosen": -0.283203125,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2466,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -6.5,
+      "step": 384
+    },
+    {
+      "epoch": 0.4723926380368098,
+      "grad_norm": 40.48440514462701,
+      "learning_rate": 6.809917355371901e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3252,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.2578125,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -5.78125,
+      "step": 385
+    },
+    {
+      "epoch": 0.4736196319018405,
+      "grad_norm": 40.628340663991764,
+      "learning_rate": 6.806611570247933e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3025,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -0.359375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -6.0625,
+      "step": 386
+    },
+    {
+      "epoch": 0.47484662576687114,
+      "grad_norm": 51.72741620856106,
+      "learning_rate": 6.803305785123966e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.90625,
+      "step": 387
+    },
+    {
+      "epoch": 0.47607361963190187,
+      "grad_norm": 43.50814776279542,
+      "learning_rate": 6.8e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3977,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.05029296875,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -5.09375,
+      "step": 388
+    },
+    {
+      "epoch": 0.47730061349693254,
+      "grad_norm": 44.9083628878753,
+      "learning_rate": 6.796694214876033e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3913,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.44140625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.625,
+      "step": 389
+    },
+    {
+      "epoch": 0.4785276073619632,
+      "grad_norm": 42.893730022875914,
+      "learning_rate": 6.793388429752066e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3759,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": -0.263671875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -6.0,
+      "step": 390
+    },
+    {
+      "epoch": 0.47975460122699387,
+      "grad_norm": 46.56902163562806,
+      "learning_rate": 6.790082644628099e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4122,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.3984375,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.5625,
+      "step": 391
+    },
+    {
+      "epoch": 0.48098159509202454,
+      "grad_norm": 32.46380915314015,
+      "learning_rate": 6.786776859504132e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2695,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.65625,
+      "step": 392
+    },
+    {
+      "epoch": 0.4822085889570552,
+      "grad_norm": 36.411647842869584,
+      "learning_rate": 6.783471074380165e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.27734375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2936,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.66015625,
+      "rewards/margins": 6.125,
+      "rewards/rejected": -5.46875,
+      "step": 393
+    },
+    {
+      "epoch": 0.4834355828220859,
+      "grad_norm": 36.43693693488016,
+      "learning_rate": 6.780165289256199e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2504,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.7265625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.34375,
+      "step": 394
+    },
+    {
+      "epoch": 0.48466257668711654,
+      "grad_norm": 36.92302249692751,
+      "learning_rate": 6.776859504132231e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -438.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3275,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.96875,
+      "step": 395
+    },
+    {
+      "epoch": 0.48588957055214727,
+      "grad_norm": 31.266490118113346,
+      "learning_rate": 6.773553719008264e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.486328125,
+      "logps/chosen": -338.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2215,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.462890625,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.6875,
+      "step": 396
+    },
+    {
+      "epoch": 0.48711656441717793,
+      "grad_norm": 34.937182799458625,
+      "learning_rate": 6.770247933884297e-07,
+      "logits/chosen": -0.2470703125,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2752,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.8125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.65625,
+      "step": 397
+    },
+    {
+      "epoch": 0.4883435582822086,
+      "grad_norm": 38.419527775162116,
+      "learning_rate": 6.76694214876033e-07,
+      "logits/chosen": -0.236328125,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -414.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3231,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.0625,
+      "step": 398
+    },
+    {
+      "epoch": 0.48957055214723927,
+      "grad_norm": 35.54458075122045,
+      "learning_rate": 6.763636363636363e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -330.0,
+      "loss": 0.306,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.9375,
+      "step": 399
+    },
+    {
+      "epoch": 0.49079754601226994,
+      "grad_norm": 47.751486198572856,
+      "learning_rate": 6.760330578512397e-07,
+      "logits/chosen": -0.2060546875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3896,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.90234375,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.46875,
+      "step": 400
+    },
+    {
+      "epoch": 0.4920245398773006,
+      "grad_norm": 31.519656830052252,
+      "learning_rate": 6.75702479338843e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.482421875,
+      "logps/chosen": -444.0,
+      "logps/rejected": -372.0,
+      "loss": 0.2337,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.6953125,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.34375,
+      "step": 401
+    },
+    {
+      "epoch": 0.49325153374233127,
+      "grad_norm": 36.11809592072752,
+      "learning_rate": 6.753719008264463e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3095,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.78125,
+      "step": 402
+    },
+    {
+      "epoch": 0.49447852760736194,
+      "grad_norm": 38.21909292164752,
+      "learning_rate": 6.750413223140496e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3169,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.875,
+      "step": 403
+    },
+    {
+      "epoch": 0.49570552147239266,
+      "grad_norm": 32.22630836566114,
+      "learning_rate": 6.747107438016528e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2283,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.0,
+      "step": 404
+    },
+    {
+      "epoch": 0.49693251533742333,
+      "grad_norm": 37.593793152840426,
+      "learning_rate": 6.743801652892561e-07,
+      "logits/chosen": -0.216796875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3305,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.65625,
+      "step": 405
+    },
+    {
+      "epoch": 0.498159509202454,
+      "grad_norm": 41.53333095908596,
+      "learning_rate": 6.740495867768595e-07,
+      "logits/chosen": -0.115234375,
+      "logits/rejected": -0.2109375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3932,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -4.3125,
+      "step": 406
+    },
+    {
+      "epoch": 0.49938650306748467,
+      "grad_norm": 42.42876322706536,
+      "learning_rate": 6.737190082644628e-07,
+      "logits/chosen": -0.0986328125,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3513,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 2.140625,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -4.3125,
+      "step": 407
+    },
+    {
+      "epoch": 0.5006134969325153,
+      "grad_norm": 48.89324534838805,
+      "learning_rate": 6.733884297520661e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -418.0,
+      "logps/rejected": -348.0,
+      "loss": 0.4164,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 4.78125,
+      "rewards/rejected": -3.984375,
+      "step": 408
+    },
+    {
+      "epoch": 0.501840490797546,
+      "grad_norm": 45.10686444580896,
+      "learning_rate": 6.730578512396694e-07,
+      "logits/chosen": -0.1318359375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -416.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3079,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.46875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -4.75,
+      "step": 409
+    },
+    {
+      "epoch": 0.5030674846625767,
+      "grad_norm": 38.62243714595701,
+      "learning_rate": 6.727272727272727e-07,
+      "logits/chosen": -0.119140625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2966,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.3125,
+      "step": 410
+    },
+    {
+      "epoch": 0.5042944785276073,
+      "grad_norm": 37.41136253841519,
+      "learning_rate": 6.72396694214876e-07,
+      "logits/chosen": -0.11083984375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -306.0,
+      "loss": 0.284,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.76953125,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.9375,
+      "step": 411
+    },
+    {
+      "epoch": 0.505521472392638,
+      "grad_norm": 47.69531455944101,
+      "learning_rate": 6.720661157024794e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3619,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 5.3125,
+      "rewards/rejected": -4.28125,
+      "step": 412
+    },
+    {
+      "epoch": 0.5067484662576687,
+      "grad_norm": 39.571364824526874,
+      "learning_rate": 6.717355371900826e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3069,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.59375,
+      "step": 413
+    },
+    {
+      "epoch": 0.5079754601226993,
+      "grad_norm": 34.32054957685153,
+      "learning_rate": 6.714049586776859e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2497,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.8125,
+      "step": 414
+    },
+    {
+      "epoch": 0.50920245398773,
+      "grad_norm": 35.2489180238823,
+      "learning_rate": 6.710743801652892e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -346.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2788,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.84375,
+      "step": 415
+    },
+    {
+      "epoch": 0.5104294478527608,
+      "grad_norm": 28.41248357906051,
+      "learning_rate": 6.707438016528925e-07,
+      "logits/chosen": -0.130859375,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -310.0,
+      "loss": 0.186,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 0.8515625,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.75,
+      "step": 416
+    },
+    {
+      "epoch": 0.5116564417177915,
+      "grad_norm": 35.52909474934163,
+      "learning_rate": 6.704132231404958e-07,
+      "logits/chosen": -0.328125,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -434.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2995,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -5.53125,
+      "step": 417
+    },
+    {
+      "epoch": 0.5128834355828221,
+      "grad_norm": 36.29519060781132,
+      "learning_rate": 6.700826446280992e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -292.0,
+      "loss": 0.2828,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.09375,
+      "step": 418
+    },
+    {
+      "epoch": 0.5141104294478528,
+      "grad_norm": 31.87680208130441,
+      "learning_rate": 6.697520661157025e-07,
+      "logits/chosen": -0.181640625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2486,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.59375,
+      "step": 419
+    },
+    {
+      "epoch": 0.5153374233128835,
+      "grad_norm": 40.881697279170936,
+      "learning_rate": 6.694214876033058e-07,
+      "logits/chosen": -0.1162109375,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -344.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3244,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.75,
+      "step": 420
+    },
+    {
+      "epoch": 0.5165644171779141,
+      "grad_norm": 34.11082321167536,
+      "learning_rate": 6.69090909090909e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -312.0,
+      "loss": 0.339,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.40625,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -4.8125,
+      "step": 421
+    },
+    {
+      "epoch": 0.5177914110429448,
+      "grad_norm": 35.18821029472038,
+      "learning_rate": 6.687603305785123e-07,
+      "logits/chosen": -0.0810546875,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3166,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.34375,
+      "step": 422
+    },
+    {
+      "epoch": 0.5190184049079755,
+      "grad_norm": 41.06855790218669,
+      "learning_rate": 6.684297520661156e-07,
+      "logits/chosen": -0.1982421875,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3606,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.6640625,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.96875,
+      "step": 423
+    },
+    {
+      "epoch": 0.5202453987730061,
+      "grad_norm": 36.724024878608596,
+      "learning_rate": 6.68099173553719e-07,
+      "logits/chosen": -0.123046875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3438,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.5625,
+      "step": 424
+    },
+    {
+      "epoch": 0.5214723926380368,
+      "grad_norm": 41.01181618114964,
+      "learning_rate": 6.677685950413223e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -342.0,
+      "loss": 0.4073,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.3203125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.96875,
+      "step": 425
+    },
+    {
+      "epoch": 0.5226993865030675,
+      "grad_norm": 44.64404869891064,
+      "learning_rate": 6.674380165289256e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3035,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.46875,
+      "step": 426
+    },
+    {
+      "epoch": 0.5239263803680981,
+      "grad_norm": 91.8646628453638,
+      "learning_rate": 6.671074380165289e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -328.0,
+      "loss": 0.299,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.09375,
+      "step": 427
+    },
+    {
+      "epoch": 0.5251533742331288,
+      "grad_norm": 36.99313104379447,
+      "learning_rate": 6.667768595041322e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3507,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.66015625,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -5.21875,
+      "step": 428
+    },
+    {
+      "epoch": 0.5263803680981595,
+      "grad_norm": 33.0190569729192,
+      "learning_rate": 6.664462809917356e-07,
+      "logits/chosen": -0.19140625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -360.0,
+      "logps/rejected": -366.0,
+      "loss": 0.246,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.443359375,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -6.46875,
+      "step": 429
+    },
+    {
+      "epoch": 0.5276073619631901,
+      "grad_norm": 30.408983999620382,
+      "learning_rate": 6.661157024793388e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2248,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.9921875,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.84375,
+      "step": 430
+    },
+    {
+      "epoch": 0.5288343558282208,
+      "grad_norm": 47.87075432548269,
+      "learning_rate": 6.657851239669421e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -336.0,
+      "loss": 0.4545,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.45703125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -5.25,
+      "step": 431
+    },
+    {
+      "epoch": 0.5300613496932516,
+      "grad_norm": 43.65339869428465,
+      "learning_rate": 6.654545454545454e-07,
+      "logits/chosen": -0.0927734375,
+      "logits/rejected": -0.1435546875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3724,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -4.875,
+      "step": 432
+    },
+    {
+      "epoch": 0.5312883435582823,
+      "grad_norm": 45.133784704459096,
+      "learning_rate": 6.651239669421487e-07,
+      "logits/chosen": -0.078125,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -308.0,
+      "loss": 0.3416,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.39453125,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.34375,
+      "step": 433
+    },
+    {
+      "epoch": 0.5325153374233129,
+      "grad_norm": 40.47930492236747,
+      "learning_rate": 6.64793388429752e-07,
+      "logits/chosen": -0.232421875,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -346.0,
+      "loss": 0.254,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.53125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.25,
+      "step": 434
+    },
+    {
+      "epoch": 0.5337423312883436,
+      "grad_norm": 41.262979059220854,
+      "learning_rate": 6.644628099173554e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3119,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.40625,
+      "step": 435
+    },
+    {
+      "epoch": 0.5349693251533743,
+      "grad_norm": 38.67538014065472,
+      "learning_rate": 6.641322314049587e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2467,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 7.15625,
+      "rewards/rejected": -5.46875,
+      "step": 436
+    },
+    {
+      "epoch": 0.5361963190184049,
+      "grad_norm": 31.32861125197491,
+      "learning_rate": 6.63801652892562e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -338.0,
+      "loss": 0.248,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 6.65625,
+      "rewards/rejected": -5.5625,
+      "step": 437
+    },
+    {
+      "epoch": 0.5374233128834356,
+      "grad_norm": 42.810824724719026,
+      "learning_rate": 6.634710743801653e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.2392578125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3969,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 5.1875,
+      "rewards/rejected": -4.375,
+      "step": 438
+    },
+    {
+      "epoch": 0.5386503067484663,
+      "grad_norm": 36.79651022084742,
+      "learning_rate": 6.631404958677685e-07,
+      "logits/chosen": -0.083984375,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.1875,
+      "step": 439
+    },
+    {
+      "epoch": 0.5398773006134969,
+      "grad_norm": 49.84881532208057,
+      "learning_rate": 6.628099173553718e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.2333984375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3608,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.384765625,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -4.6875,
+      "step": 440
+    },
+    {
+      "epoch": 0.5411042944785276,
+      "grad_norm": 37.976634967176224,
+      "learning_rate": 6.624793388429752e-07,
+      "logits/chosen": -0.06640625,
+      "logits/rejected": -0.1826171875,
+      "logps/chosen": -332.0,
+      "logps/rejected": -300.0,
+      "loss": 0.328,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.75,
+      "step": 441
+    },
+    {
+      "epoch": 0.5423312883435583,
+      "grad_norm": 37.80188895923485,
+      "learning_rate": 6.621487603305785e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3304,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.8125,
+      "step": 442
+    },
+    {
+      "epoch": 0.5435582822085889,
+      "grad_norm": 40.24381216440989,
+      "learning_rate": 6.618181818181818e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3159,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.4140625,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.59375,
+      "step": 443
+    },
+    {
+      "epoch": 0.5447852760736196,
+      "grad_norm": 37.922945446281,
+      "learning_rate": 6.614876033057851e-07,
+      "logits/chosen": -0.201171875,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -446.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2446,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.09375,
+      "rewards/margins": 7.25,
+      "rewards/rejected": -5.15625,
+      "step": 444
+    },
+    {
+      "epoch": 0.5460122699386503,
+      "grad_norm": 38.45325276499507,
+      "learning_rate": 6.611570247933884e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -300.0,
+      "loss": 0.389,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.125,
+      "step": 445
+    },
+    {
+      "epoch": 0.5472392638036809,
+      "grad_norm": 33.02180393783636,
+      "learning_rate": 6.608264462809917e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2348,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.46875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.25,
+      "step": 446
+    },
+    {
+      "epoch": 0.5484662576687117,
+      "grad_norm": 31.670625439038492,
+      "learning_rate": 6.604958677685951e-07,
+      "logits/chosen": -0.1767578125,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2551,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.875,
+      "rewards/margins": 7.15625,
+      "rewards/rejected": -5.28125,
+      "step": 447
+    },
+    {
+      "epoch": 0.5496932515337424,
+      "grad_norm": 37.605046960174505,
+      "learning_rate": 6.601652892561983e-07,
+      "logits/chosen": -0.10888671875,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3209,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.90234375,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -4.25,
+      "step": 448
+    },
+    {
+      "epoch": 0.550920245398773,
+      "grad_norm": 34.080535896094865,
+      "learning_rate": 6.598347107438016e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2712,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9453125,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -4.53125,
+      "step": 449
+    },
+    {
+      "epoch": 0.5521472392638037,
+      "grad_norm": 35.069217343013015,
+      "learning_rate": 6.595041322314049e-07,
+      "logits/chosen": -0.255859375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -288.0,
+      "loss": 0.285,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.625,
+      "step": 450
+    },
+    {
+      "epoch": 0.5533742331288344,
+      "grad_norm": 33.87966228532685,
+      "learning_rate": 6.591735537190082e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -340.0,
+      "logps/rejected": -280.0,
+      "loss": 0.3043,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -4.5,
+      "step": 451
+    },
+    {
+      "epoch": 0.554601226993865,
+      "grad_norm": 35.08833886262592,
+      "learning_rate": 6.588429752066115e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -454.0,
+      "logps/rejected": -328.0,
+      "loss": 0.258,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.421875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.21875,
+      "step": 452
+    },
+    {
+      "epoch": 0.5558282208588957,
+      "grad_norm": 30.509116856340853,
+      "learning_rate": 6.585123966942149e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.48828125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -366.0,
+      "loss": 0.214,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 2.578125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -4.78125,
+      "step": 453
+    },
+    {
+      "epoch": 0.5570552147239264,
+      "grad_norm": 36.06945734305074,
+      "learning_rate": 6.581818181818182e-07,
+      "logits/chosen": -0.11572265625,
+      "logits/rejected": -0.19921875,
+      "logps/chosen": -338.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2945,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.859375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.625,
+      "step": 454
+    },
+    {
+      "epoch": 0.558282208588957,
+      "grad_norm": 35.814869576892626,
+      "learning_rate": 6.578512396694215e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.36328125,
+      "logps/chosen": -366.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2399,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.9609375,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -4.8125,
+      "step": 455
+    },
+    {
+      "epoch": 0.5595092024539877,
+      "grad_norm": 38.450506141939925,
+      "learning_rate": 6.575206611570248e-07,
+      "logits/chosen": -0.12060546875,
+      "logits/rejected": -0.2236328125,
+      "logps/chosen": -312.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2952,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.125,
+      "step": 456
+    },
+    {
+      "epoch": 0.5607361963190184,
+      "grad_norm": 37.93391723235656,
+      "learning_rate": 6.57190082644628e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3025,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -4.5625,
+      "step": 457
+    },
+    {
+      "epoch": 0.561963190184049,
+      "grad_norm": 33.44669035850576,
+      "learning_rate": 6.568595041322313e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -370.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2297,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.75,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.125,
+      "step": 458
+    },
+    {
+      "epoch": 0.5631901840490797,
+      "grad_norm": 34.06851635286411,
+      "learning_rate": 6.565289256198347e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2786,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.0625,
+      "step": 459
+    },
+    {
+      "epoch": 0.5644171779141104,
+      "grad_norm": 34.68751260016757,
+      "learning_rate": 6.56198347107438e-07,
+      "logits/chosen": -0.1630859375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -330.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.76171875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -6.03125,
+      "step": 460
+    },
+    {
+      "epoch": 0.5656441717791411,
+      "grad_norm": 38.71078681224242,
+      "learning_rate": 6.558677685950413e-07,
+      "logits/chosen": -0.3515625,
+      "logits/rejected": -0.48046875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2741,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.80078125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.34375,
+      "step": 461
+    },
+    {
+      "epoch": 0.5668711656441717,
+      "grad_norm": 39.327400074833584,
+      "learning_rate": 6.555371900826446e-07,
+      "logits/chosen": -0.11279296875,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3248,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.875,
+      "step": 462
+    },
+    {
+      "epoch": 0.5680981595092025,
+      "grad_norm": 37.637164385131356,
+      "learning_rate": 6.552066115702479e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -364.0,
+      "logps/rejected": -360.0,
+      "loss": 0.2848,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.98046875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.46875,
+      "step": 463
+    },
+    {
+      "epoch": 0.5693251533742332,
+      "grad_norm": 41.25059009443085,
+      "learning_rate": 6.548760330578512e-07,
+      "logits/chosen": -0.09619140625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2951,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.8671875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.34375,
+      "step": 464
+    },
+    {
+      "epoch": 0.5705521472392638,
+      "grad_norm": 27.065991136479923,
+      "learning_rate": 6.545454545454546e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -346.0,
+      "loss": 0.1759,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -6.0625,
+      "step": 465
+    },
+    {
+      "epoch": 0.5717791411042945,
+      "grad_norm": 43.40449571309543,
+      "learning_rate": 6.542148760330578e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2899,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.9375,
+      "step": 466
+    },
+    {
+      "epoch": 0.5730061349693252,
+      "grad_norm": 40.392254031751044,
+      "learning_rate": 6.538842975206611e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -384.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3087,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.2890625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.28125,
+      "step": 467
+    },
+    {
+      "epoch": 0.5742331288343558,
+      "grad_norm": 41.32661857185299,
+      "learning_rate": 6.535537190082644e-07,
+      "logits/chosen": -0.171875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3022,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.78125,
+      "step": 468
+    },
+    {
+      "epoch": 0.5754601226993865,
+      "grad_norm": 29.07874137818143,
+      "learning_rate": 6.532231404958677e-07,
+      "logits/chosen": -0.12451171875,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -306.0,
+      "loss": 0.1838,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.6875,
+      "step": 469
+    },
+    {
+      "epoch": 0.5766871165644172,
+      "grad_norm": 35.40917214666752,
+      "learning_rate": 6.52892561983471e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.2353515625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -324.0,
+      "loss": 0.301,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.34375,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -5.40625,
+      "step": 470
+    },
+    {
+      "epoch": 0.5779141104294478,
+      "grad_norm": 38.07545863951338,
+      "learning_rate": 6.525619834710744e-07,
+      "logits/chosen": -0.03369140625,
+      "logits/rejected": -0.203125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3263,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -5.125,
+      "step": 471
+    },
+    {
+      "epoch": 0.5791411042944785,
+      "grad_norm": 40.21918996431336,
+      "learning_rate": 6.522314049586777e-07,
+      "logits/chosen": -0.08447265625,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -406.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3173,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -5.15625,
+      "step": 472
+    },
+    {
+      "epoch": 0.5803680981595092,
+      "grad_norm": 37.03158250739783,
+      "learning_rate": 6.51900826446281e-07,
+      "logits/chosen": -0.25,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -378.0,
+      "loss": 0.3328,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.90625,
+      "step": 473
+    },
+    {
+      "epoch": 0.5815950920245399,
+      "grad_norm": 38.12266289452875,
+      "learning_rate": 6.515702479338843e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -430.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2648,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -5.21875,
+      "step": 474
+    },
+    {
+      "epoch": 0.5828220858895705,
+      "grad_norm": 31.08159067757818,
+      "learning_rate": 6.512396694214875e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.4140625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2361,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 7.59375,
+      "rewards/rejected": -5.59375,
+      "step": 475
+    },
+    {
+      "epoch": 0.5840490797546012,
+      "grad_norm": 31.0829148956811,
+      "learning_rate": 6.509090909090908e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2813,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.185546875,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -5.6875,
+      "step": 476
+    },
+    {
+      "epoch": 0.5852760736196319,
+      "grad_norm": 36.7408304979899,
+      "learning_rate": 6.505785123966942e-07,
+      "logits/chosen": -0.244140625,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -354.0,
+      "loss": 0.257,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.75,
+      "step": 477
+    },
+    {
+      "epoch": 0.5865030674846625,
+      "grad_norm": 34.872893059723594,
+      "learning_rate": 6.502479338842975e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2988,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.84375,
+      "step": 478
+    },
+    {
+      "epoch": 0.5877300613496933,
+      "grad_norm": 35.26754811760802,
+      "learning_rate": 6.499173553719008e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2784,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.240234375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.96875,
+      "step": 479
+    },
+    {
+      "epoch": 0.588957055214724,
+      "grad_norm": 33.02148763705059,
+      "learning_rate": 6.495867768595041e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2364,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.4765625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -6.125,
+      "step": 480
+    },
+    {
+      "epoch": 0.5901840490797546,
+      "grad_norm": 39.97994667673503,
+      "learning_rate": 6.492561983471074e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -378.0,
+      "loss": 0.3732,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.84765625,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.46875,
+      "step": 481
+    },
+    {
+      "epoch": 0.5914110429447853,
+      "grad_norm": 33.72401411710678,
+      "learning_rate": 6.489256198347108e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2739,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.4375,
+      "step": 482
+    },
+    {
+      "epoch": 0.592638036809816,
+      "grad_norm": 34.19291475160813,
+      "learning_rate": 6.485950413223141e-07,
+      "logits/chosen": -0.0703125,
+      "logits/rejected": -0.2138671875,
+      "logps/chosen": -346.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2917,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.69921875,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.5625,
+      "step": 483
+    },
+    {
+      "epoch": 0.5938650306748466,
+      "grad_norm": 35.398942571150606,
+      "learning_rate": 6.482644628099173e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3155,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.1875,
+      "rewards/rejected": -4.875,
+      "step": 484
+    },
+    {
+      "epoch": 0.5950920245398773,
+      "grad_norm": 30.26872146503766,
+      "learning_rate": 6.479338842975206e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -306.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.625,
+      "step": 485
+    },
+    {
+      "epoch": 0.596319018404908,
+      "grad_norm": 35.38748335338782,
+      "learning_rate": 6.476033057851239e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2658,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -4.96875,
+      "step": 486
+    },
+    {
+      "epoch": 0.5975460122699386,
+      "grad_norm": 30.285435869887902,
+      "learning_rate": 6.472727272727272e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -350.0,
+      "loss": 0.233,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.5625,
+      "step": 487
+    },
+    {
+      "epoch": 0.5987730061349693,
+      "grad_norm": 37.89908939404138,
+      "learning_rate": 6.469421487603306e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -416.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2436,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.0625,
+      "step": 488
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 30.515429632482274,
+      "learning_rate": 6.466115702479339e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -326.0,
+      "loss": 0.272,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 7.3125,
+      "rewards/rejected": -5.5,
+      "step": 489
+    },
+    {
+      "epoch": 0.6012269938650306,
+      "grad_norm": 40.37971138033072,
+      "learning_rate": 6.462809917355372e-07,
+      "logits/chosen": -0.09765625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3004,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -5.1875,
+      "step": 490
+    },
+    {
+      "epoch": 0.6024539877300613,
+      "grad_norm": 34.45143017996783,
+      "learning_rate": 6.459504132231405e-07,
+      "logits/chosen": -0.08203125,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2549,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.4375,
+      "step": 491
+    },
+    {
+      "epoch": 0.603680981595092,
+      "grad_norm": 36.58265725949472,
+      "learning_rate": 6.456198347107438e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -378.0,
+      "loss": 0.2737,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.9375,
+      "step": 492
+    },
+    {
+      "epoch": 0.6049079754601226,
+      "grad_norm": 42.46481332966053,
+      "learning_rate": 6.45289256198347e-07,
+      "logits/chosen": -0.024169921875,
+      "logits/rejected": -0.134765625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3917,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.28125,
+      "rewards/rejected": -4.25,
+      "step": 493
+    },
+    {
+      "epoch": 0.6061349693251534,
+      "grad_norm": 33.59404853921049,
+      "learning_rate": 6.449586776859504e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -422.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2643,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.09375,
+      "step": 494
+    },
+    {
+      "epoch": 0.6073619631901841,
+      "grad_norm": 37.484469870256234,
+      "learning_rate": 6.446280991735537e-07,
+      "logits/chosen": -0.1220703125,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3066,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.21875,
+      "step": 495
+    },
+    {
+      "epoch": 0.6085889570552148,
+      "grad_norm": 40.63137683640751,
+      "learning_rate": 6.44297520661157e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3177,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -5.28125,
+      "step": 496
+    },
+    {
+      "epoch": 0.6098159509202454,
+      "grad_norm": 41.064298986990785,
+      "learning_rate": 6.439669421487603e-07,
+      "logits/chosen": -0.12158203125,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3865,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.85546875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.5,
+      "step": 497
+    },
+    {
+      "epoch": 0.6110429447852761,
+      "grad_norm": 34.6148352569408,
+      "learning_rate": 6.436363636363636e-07,
+      "logits/chosen": -0.2431640625,
+      "logits/rejected": -0.421875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -358.0,
+      "loss": 0.2817,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.65625,
+      "step": 498
+    },
+    {
+      "epoch": 0.6122699386503068,
+      "grad_norm": 36.60546803390198,
+      "learning_rate": 6.433057851239669e-07,
+      "logits/chosen": -0.115234375,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2842,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.125,
+      "step": 499
+    },
+    {
+      "epoch": 0.6134969325153374,
+      "grad_norm": 33.21675862161022,
+      "learning_rate": 6.429752066115703e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2718,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.21875,
+      "step": 500
+    },
+    {
+      "epoch": 0.6147239263803681,
+      "grad_norm": 32.544310987499145,
+      "learning_rate": 6.426446280991735e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -344.0,
+      "loss": 0.26,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.625,
+      "step": 501
+    },
+    {
+      "epoch": 0.6159509202453988,
+      "grad_norm": 32.09170254227315,
+      "learning_rate": 6.423140495867768e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.4296875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2639,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.515625,
+      "rewards/margins": 7.46875,
+      "rewards/rejected": -5.96875,
+      "step": 502
+    },
+    {
+      "epoch": 0.6171779141104294,
+      "grad_norm": 35.53103605940553,
+      "learning_rate": 6.419834710743801e-07,
+      "logits/chosen": 0.01153564453125,
+      "logits/rejected": -0.1904296875,
+      "logps/chosen": -326.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2795,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.25,
+      "step": 503
+    },
+    {
+      "epoch": 0.6184049079754601,
+      "grad_norm": 30.299429390734634,
+      "learning_rate": 6.416528925619834e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.39453125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2658,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -5.40625,
+      "step": 504
+    },
+    {
+      "epoch": 0.6196319018404908,
+      "grad_norm": 44.689012762744575,
+      "learning_rate": 6.413223140495867e-07,
+      "logits/chosen": -0.220703125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -394.0,
+      "loss": 0.3776,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.369140625,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -6.15625,
+      "step": 505
+    },
+    {
+      "epoch": 0.6208588957055214,
+      "grad_norm": 32.33588195685234,
+      "learning_rate": 6.409917355371901e-07,
+      "logits/chosen": -0.08349609375,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2588,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.90625,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.875,
+      "step": 506
+    },
+    {
+      "epoch": 0.6220858895705521,
+      "grad_norm": 29.030680879470466,
+      "learning_rate": 6.406611570247934e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -308.0,
+      "loss": 0.1953,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.90234375,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -6.15625,
+      "step": 507
+    },
+    {
+      "epoch": 0.6233128834355828,
+      "grad_norm": 31.4626526695897,
+      "learning_rate": 6.403305785123967e-07,
+      "logits/chosen": -0.0201416015625,
+      "logits/rejected": -0.2021484375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -310.0,
+      "loss": 0.24,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.671875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.59375,
+      "step": 508
+    },
+    {
+      "epoch": 0.6245398773006134,
+      "grad_norm": 42.61520515559976,
+      "learning_rate": 6.4e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.40234375,
+      "logps/chosen": -466.0,
+      "logps/rejected": -384.0,
+      "loss": 0.3973,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.8984375,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -6.0625,
+      "step": 509
+    },
+    {
+      "epoch": 0.6257668711656442,
+      "grad_norm": 30.165060750168916,
+      "learning_rate": 6.396694214876032e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2226,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.306640625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.625,
+      "step": 510
+    },
+    {
+      "epoch": 0.6269938650306749,
+      "grad_norm": 40.15670022385489,
+      "learning_rate": 6.393388429752065e-07,
+      "logits/chosen": -0.271484375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3348,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.81640625,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -6.1875,
+      "step": 511
+    },
+    {
+      "epoch": 0.6282208588957056,
+      "grad_norm": 44.290408519253646,
+      "learning_rate": 6.390082644628099e-07,
+      "logits/chosen": -0.09130859375,
+      "logits/rejected": -0.2431640625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -384.0,
+      "loss": 0.3621,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.35546875,
+      "rewards/margins": 5.375,
+      "rewards/rejected": -5.0,
+      "step": 512
+    },
+    {
+      "epoch": 0.6294478527607362,
+      "grad_norm": 37.44250565144745,
+      "learning_rate": 6.386776859504132e-07,
+      "logits/chosen": -0.09521484375,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -312.0,
+      "loss": 0.337,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -5.375,
+      "step": 513
+    },
+    {
+      "epoch": 0.6306748466257669,
+      "grad_norm": 29.97245642529846,
+      "learning_rate": 6.383471074380165e-07,
+      "logits/chosen": -0.0830078125,
+      "logits/rejected": -0.2412109375,
+      "logps/chosen": -330.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2436,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": -0.051513671875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -6.03125,
+      "step": 514
+    },
+    {
+      "epoch": 0.6319018404907976,
+      "grad_norm": 36.989875573277935,
+      "learning_rate": 6.380165289256198e-07,
+      "logits/chosen": -0.09912109375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -348.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2882,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": -0.041748046875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.84375,
+      "step": 515
+    },
+    {
+      "epoch": 0.6331288343558282,
+      "grad_norm": 31.306859511596812,
+      "learning_rate": 6.376859504132231e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2432,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.1142578125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.84375,
+      "step": 516
+    },
+    {
+      "epoch": 0.6343558282208589,
+      "grad_norm": 43.67170294973452,
+      "learning_rate": 6.373553719008264e-07,
+      "logits/chosen": -0.12255859375,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3884,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.53125,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -5.03125,
+      "step": 517
+    },
+    {
+      "epoch": 0.6355828220858896,
+      "grad_norm": 29.354877414703584,
+      "learning_rate": 6.370247933884298e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -342.0,
+      "loss": 0.1963,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.34375,
+      "step": 518
+    },
+    {
+      "epoch": 0.6368098159509202,
+      "grad_norm": 40.61854272265548,
+      "learning_rate": 6.36694214876033e-07,
+      "logits/chosen": -0.197265625,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -334.0,
+      "loss": 0.299,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.75390625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.5,
+      "step": 519
+    },
+    {
+      "epoch": 0.6380368098159509,
+      "grad_norm": 38.052325639181014,
+      "learning_rate": 6.363636363636363e-07,
+      "logits/chosen": -0.228515625,
+      "logits/rejected": -0.46484375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3436,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.6875,
+      "step": 520
+    },
+    {
+      "epoch": 0.6392638036809816,
+      "grad_norm": 32.37002227141111,
+      "learning_rate": 6.360330578512396e-07,
+      "logits/chosen": -0.03271484375,
+      "logits/rejected": -0.171875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2525,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.5,
+      "step": 521
+    },
+    {
+      "epoch": 0.6404907975460122,
+      "grad_norm": 40.41274539784558,
+      "learning_rate": 6.357024793388429e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -420.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3152,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.96875,
+      "step": 522
+    },
+    {
+      "epoch": 0.6417177914110429,
+      "grad_norm": 42.30362208986331,
+      "learning_rate": 6.353719008264462e-07,
+      "logits/chosen": -0.16015625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3605,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.6171875,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -5.03125,
+      "step": 523
+    },
+    {
+      "epoch": 0.6429447852760736,
+      "grad_norm": 46.893532525519284,
+      "learning_rate": 6.350413223140496e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -420.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3361,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.0625,
+      "step": 524
+    },
+    {
+      "epoch": 0.6441717791411042,
+      "grad_norm": 49.422180518954264,
+      "learning_rate": 6.347107438016529e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.185546875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3927,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.498046875,
+      "rewards/margins": 5.15625,
+      "rewards/rejected": -4.65625,
+      "step": 525
+    },
+    {
+      "epoch": 0.645398773006135,
+      "grad_norm": 38.5530601675623,
+      "learning_rate": 6.343801652892562e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2504,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -5.5625,
+      "step": 526
+    },
+    {
+      "epoch": 0.6466257668711657,
+      "grad_norm": 35.94650372227145,
+      "learning_rate": 6.340495867768595e-07,
+      "logits/chosen": -0.0859375,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2888,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.9609375,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.1875,
+      "step": 527
+    },
+    {
+      "epoch": 0.6478527607361964,
+      "grad_norm": 33.57585022481078,
+      "learning_rate": 6.337190082644627e-07,
+      "logits/chosen": -0.2099609375,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2986,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.71484375,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.71875,
+      "step": 528
+    },
+    {
+      "epoch": 0.649079754601227,
+      "grad_norm": 35.863933883867006,
+      "learning_rate": 6.33388429752066e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3657,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.0390625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.1875,
+      "step": 529
+    },
+    {
+      "epoch": 0.6503067484662577,
+      "grad_norm": 41.86945425639466,
+      "learning_rate": 6.330578512396694e-07,
+      "logits/chosen": -0.12353515625,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3197,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.375,
+      "step": 530
+    },
+    {
+      "epoch": 0.6515337423312884,
+      "grad_norm": 33.3934976525138,
+      "learning_rate": 6.327272727272727e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2751,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -5.1875,
+      "step": 531
+    },
+    {
+      "epoch": 0.652760736196319,
+      "grad_norm": 35.72821914002383,
+      "learning_rate": 6.32396694214876e-07,
+      "logits/chosen": -0.07958984375,
+      "logits/rejected": -0.2197265625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2782,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.875,
+      "step": 532
+    },
+    {
+      "epoch": 0.6539877300613497,
+      "grad_norm": 36.05732236551947,
+      "learning_rate": 6.320661157024793e-07,
+      "logits/chosen": -0.2421875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -402.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2611,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.8671875,
+      "rewards/margins": 7.3125,
+      "rewards/rejected": -5.4375,
+      "step": 533
+    },
+    {
+      "epoch": 0.6552147239263804,
+      "grad_norm": 43.004226294889776,
+      "learning_rate": 6.317355371900826e-07,
+      "logits/chosen": -0.10302734375,
+      "logits/rejected": -0.228515625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2813,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -4.84375,
+      "step": 534
+    },
+    {
+      "epoch": 0.656441717791411,
+      "grad_norm": 39.5682500491772,
+      "learning_rate": 6.31404958677686e-07,
+      "logits/chosen": -0.2109375,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2961,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.380859375,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.5625,
+      "step": 535
+    },
+    {
+      "epoch": 0.6576687116564417,
+      "grad_norm": 36.55745143954986,
+      "learning_rate": 6.310743801652893e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2275,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.4375,
+      "step": 536
+    },
+    {
+      "epoch": 0.6588957055214724,
+      "grad_norm": 45.10266990819029,
+      "learning_rate": 6.307438016528925e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -378.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3939,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.265625,
+      "rewards/margins": 5.125,
+      "rewards/rejected": -4.84375,
+      "step": 537
+    },
+    {
+      "epoch": 0.660122699386503,
+      "grad_norm": 44.09458580737532,
+      "learning_rate": 6.304132231404958e-07,
+      "logits/chosen": -0.1123046875,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3783,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.84375,
+      "step": 538
+    },
+    {
+      "epoch": 0.6613496932515337,
+      "grad_norm": 36.16672414551654,
+      "learning_rate": 6.300826446280991e-07,
+      "logits/chosen": -0.11181640625,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -340.0,
+      "loss": 0.309,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.40625,
+      "step": 539
+    },
+    {
+      "epoch": 0.6625766871165644,
+      "grad_norm": 39.077710779231424,
+      "learning_rate": 6.297520661157024e-07,
+      "logits/chosen": -0.0908203125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3217,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4921875,
+      "rewards/margins": 6.65625,
+      "rewards/rejected": -5.15625,
+      "step": 540
+    },
+    {
+      "epoch": 0.6638036809815951,
+      "grad_norm": 35.971169171679556,
+      "learning_rate": 6.294214876033058e-07,
+      "logits/chosen": -0.10986328125,
+      "logits/rejected": -0.2080078125,
+      "logps/chosen": -386.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2834,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.75390625,
+      "rewards/margins": 5.5625,
+      "rewards/rejected": -4.8125,
+      "step": 541
+    },
+    {
+      "epoch": 0.6650306748466258,
+      "grad_norm": 34.168657597082216,
+      "learning_rate": 6.290909090909091e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2812,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.65625,
+      "step": 542
+    },
+    {
+      "epoch": 0.6662576687116565,
+      "grad_norm": 36.23729440291612,
+      "learning_rate": 6.287603305785124e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -444.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2622,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -4.78125,
+      "step": 543
+    },
+    {
+      "epoch": 0.6674846625766871,
+      "grad_norm": 45.238447000820585,
+      "learning_rate": 6.284297520661157e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -374.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3695,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.91015625,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.25,
+      "step": 544
+    },
+    {
+      "epoch": 0.6687116564417178,
+      "grad_norm": 43.061817496062964,
+      "learning_rate": 6.28099173553719e-07,
+      "logits/chosen": -0.2138671875,
+      "logits/rejected": -0.388671875,
+      "logps/chosen": -452.0,
+      "logps/rejected": -360.0,
+      "loss": 0.4047,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -4.5625,
+      "step": 545
+    },
+    {
+      "epoch": 0.6699386503067485,
+      "grad_norm": 40.8862609592416,
+      "learning_rate": 6.277685950413222e-07,
+      "logits/chosen": -0.10205078125,
+      "logits/rejected": -0.20703125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3402,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.1953125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.09375,
+      "step": 546
+    },
+    {
+      "epoch": 0.6711656441717792,
+      "grad_norm": 32.99712535949366,
+      "learning_rate": 6.274380165289256e-07,
+      "logits/chosen": -0.1591796875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2883,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.9375,
+      "step": 547
+    },
+    {
+      "epoch": 0.6723926380368098,
+      "grad_norm": 40.33054062017028,
+      "learning_rate": 6.271074380165289e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2583,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.71875,
+      "step": 548
+    },
+    {
+      "epoch": 0.6736196319018405,
+      "grad_norm": 37.72507019090557,
+      "learning_rate": 6.267768595041322e-07,
+      "logits/chosen": -0.11474609375,
+      "logits/rejected": -0.330078125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.84375,
+      "step": 549
+    },
+    {
+      "epoch": 0.6748466257668712,
+      "grad_norm": 37.30173963897267,
+      "learning_rate": 6.264462809917355e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3323,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.59375,
+      "step": 550
+    },
+    {
+      "epoch": 0.6760736196319018,
+      "grad_norm": 33.242769441390706,
+      "learning_rate": 6.261157024793388e-07,
+      "logits/chosen": -0.0986328125,
+      "logits/rejected": -0.2265625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2814,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.9296875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -4.59375,
+      "step": 551
+    },
+    {
+      "epoch": 0.6773006134969325,
+      "grad_norm": 35.53857403218536,
+      "learning_rate": 6.257851239669421e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3055,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -4.6875,
+      "step": 552
+    },
+    {
+      "epoch": 0.6785276073619632,
+      "grad_norm": 39.8186178694702,
+      "learning_rate": 6.254545454545455e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -358.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3556,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.60546875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.8125,
+      "step": 553
+    },
+    {
+      "epoch": 0.6797546012269938,
+      "grad_norm": 34.396570445757696,
+      "learning_rate": 6.251239669421488e-07,
+      "logits/chosen": -0.08349609375,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -424.0,
+      "logps/rejected": -326.0,
+      "loss": 0.261,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -4.6875,
+      "step": 554
+    },
+    {
+      "epoch": 0.6809815950920245,
+      "grad_norm": 37.46750845748253,
+      "learning_rate": 6.24793388429752e-07,
+      "logits/chosen": 0.0196533203125,
+      "logits/rejected": -0.107421875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3763,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.296875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.65625,
+      "step": 555
+    },
+    {
+      "epoch": 0.6822085889570552,
+      "grad_norm": 28.75095679008507,
+      "learning_rate": 6.244628099173553e-07,
+      "logits/chosen": -0.10693359375,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -414.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2294,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.0625,
+      "step": 556
+    },
+    {
+      "epoch": 0.6834355828220859,
+      "grad_norm": 32.12117772558378,
+      "learning_rate": 6.241322314049586e-07,
+      "logits/chosen": -0.11669921875,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2775,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -4.65625,
+      "step": 557
+    },
+    {
+      "epoch": 0.6846625766871166,
+      "grad_norm": 30.45940478111348,
+      "learning_rate": 6.238016528925619e-07,
+      "logits/chosen": -0.07861328125,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2442,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.8359375,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.875,
+      "step": 558
+    },
+    {
+      "epoch": 0.6858895705521473,
+      "grad_norm": 38.28759745643993,
+      "learning_rate": 6.234710743801653e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3234,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.0078125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.25,
+      "step": 559
+    },
+    {
+      "epoch": 0.6871165644171779,
+      "grad_norm": 25.47174077723704,
+      "learning_rate": 6.231404958677686e-07,
+      "logits/chosen": -0.3046875,
+      "logits/rejected": -0.451171875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2016,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -6.03125,
+      "step": 560
+    },
+    {
+      "epoch": 0.6883435582822086,
+      "grad_norm": 28.61531386213562,
+      "learning_rate": 6.228099173553719e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -300.0,
+      "loss": 0.2021,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.6875,
+      "step": 561
+    },
+    {
+      "epoch": 0.6895705521472393,
+      "grad_norm": 30.528229373220828,
+      "learning_rate": 6.224793388429752e-07,
+      "logits/chosen": -0.032470703125,
+      "logits/rejected": -0.2333984375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2462,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.5,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.5625,
+      "step": 562
+    },
+    {
+      "epoch": 0.69079754601227,
+      "grad_norm": 34.031991776353316,
+      "learning_rate": 6.221487603305785e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2904,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 7.21875,
+      "rewards/rejected": -5.65625,
+      "step": 563
+    },
+    {
+      "epoch": 0.6920245398773006,
+      "grad_norm": 24.836619890425215,
+      "learning_rate": 6.218181818181817e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -388.0,
+      "logps/rejected": -334.0,
+      "loss": 0.224,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.7109375,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.34375,
+      "step": 564
+    },
+    {
+      "epoch": 0.6932515337423313,
+      "grad_norm": 35.09286389818353,
+      "learning_rate": 6.214876033057851e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -394.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3639,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.6015625,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.84375,
+      "step": 565
+    },
+    {
+      "epoch": 0.694478527607362,
+      "grad_norm": 33.71563020260373,
+      "learning_rate": 6.211570247933884e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2462,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.1875,
+      "step": 566
+    },
+    {
+      "epoch": 0.6957055214723926,
+      "grad_norm": 25.84647564741135,
+      "learning_rate": 6.208264462809917e-07,
+      "logits/chosen": -0.053955078125,
+      "logits/rejected": -0.1875,
+      "logps/chosen": -316.0,
+      "logps/rejected": -274.0,
+      "loss": 0.2061,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.359375,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -5.09375,
+      "step": 567
+    },
+    {
+      "epoch": 0.6969325153374233,
+      "grad_norm": 37.567177147049684,
+      "learning_rate": 6.20495867768595e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3064,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.21875,
+      "step": 568
+    },
+    {
+      "epoch": 0.698159509202454,
+      "grad_norm": 32.35117698678835,
+      "learning_rate": 6.201652892561983e-07,
+      "logits/chosen": -0.10302734375,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2508,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.125,
+      "step": 569
+    },
+    {
+      "epoch": 0.6993865030674846,
+      "grad_norm": 25.32334987337483,
+      "learning_rate": 6.198347107438017e-07,
+      "logits/chosen": -0.26953125,
+      "logits/rejected": -0.404296875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -338.0,
+      "loss": 0.1649,
+      "rewards/accuracies": 0.9453125,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.34375,
+      "step": 570
+    },
+    {
+      "epoch": 0.7006134969325153,
+      "grad_norm": 31.066523168854104,
+      "learning_rate": 6.19504132231405e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -326.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2262,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.6015625,
+      "rewards/margins": 5.6875,
+      "rewards/rejected": -5.09375,
+      "step": 571
+    },
+    {
+      "epoch": 0.701840490797546,
+      "grad_norm": 32.881821449548816,
+      "learning_rate": 6.191735537190083e-07,
+      "logits/chosen": -0.1142578125,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2291,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -4.9375,
+      "step": 572
+    },
+    {
+      "epoch": 0.7030674846625767,
+      "grad_norm": 40.68529761937124,
+      "learning_rate": 6.188429752066115e-07,
+      "logits/chosen": -0.07177734375,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4034,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -5.3125,
+      "step": 573
+    },
+    {
+      "epoch": 0.7042944785276074,
+      "grad_norm": 38.820291986225,
+      "learning_rate": 6.185123966942148e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.328125,
+      "logps/chosen": -322.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2792,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.96484375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.25,
+      "step": 574
+    },
+    {
+      "epoch": 0.7055214723926381,
+      "grad_norm": 33.59717634578736,
+      "learning_rate": 6.181818181818181e-07,
+      "logits/chosen": -0.240234375,
+      "logits/rejected": -0.41015625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -370.0,
+      "loss": 0.2284,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.8203125,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.59375,
+      "step": 575
+    },
+    {
+      "epoch": 0.7067484662576687,
+      "grad_norm": 28.04916589516562,
+      "learning_rate": 6.178512396694215e-07,
+      "logits/chosen": -0.10302734375,
+      "logits/rejected": -0.3125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2001,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -5.875,
+      "step": 576
+    },
+    {
+      "epoch": 0.7079754601226994,
+      "grad_norm": 36.15869255614372,
+      "learning_rate": 6.175206611570248e-07,
+      "logits/chosen": -0.03955078125,
+      "logits/rejected": -0.1767578125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2929,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.578125,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.84375,
+      "step": 577
+    },
+    {
+      "epoch": 0.7092024539877301,
+      "grad_norm": 36.89994352340565,
+      "learning_rate": 6.171900826446281e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -334.0,
+      "loss": 0.322,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.65625,
+      "rewards/margins": 7.40625,
+      "rewards/rejected": -5.75,
+      "step": 578
+    },
+    {
+      "epoch": 0.7104294478527607,
+      "grad_norm": 42.55492037495249,
+      "learning_rate": 6.168595041322314e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3864,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.451171875,
+      "rewards/margins": 5.4375,
+      "rewards/rejected": -4.96875,
+      "step": 579
+    },
+    {
+      "epoch": 0.7116564417177914,
+      "grad_norm": 35.02868013813214,
+      "learning_rate": 6.165289256198347e-07,
+      "logits/chosen": -0.07275390625,
+      "logits/rejected": -0.2451171875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -354.0,
+      "loss": 0.229,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.73046875,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -6.25,
+      "step": 580
+    },
+    {
+      "epoch": 0.7128834355828221,
+      "grad_norm": 29.748844282456343,
+      "learning_rate": 6.16198347107438e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.37890625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2071,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.84375,
+      "step": 581
+    },
+    {
+      "epoch": 0.7141104294478527,
+      "grad_norm": 41.261123421442086,
+      "learning_rate": 6.158677685950413e-07,
+      "logits/chosen": 0.01953125,
+      "logits/rejected": -0.197265625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3132,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.9453125,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.78125,
+      "step": 582
+    },
+    {
+      "epoch": 0.7153374233128834,
+      "grad_norm": 37.14443132697755,
+      "learning_rate": 6.155371900826446e-07,
+      "logits/chosen": -0.1650390625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -382.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3036,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.875,
+      "step": 583
+    },
+    {
+      "epoch": 0.7165644171779141,
+      "grad_norm": 42.810461239623855,
+      "learning_rate": 6.152066115702479e-07,
+      "logits/chosen": -0.0693359375,
+      "logits/rejected": -0.14453125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -370.0,
+      "loss": 0.3929,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 0.5859375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.25,
+      "step": 584
+    },
+    {
+      "epoch": 0.7177914110429447,
+      "grad_norm": 44.860424166016166,
+      "learning_rate": 6.148760330578512e-07,
+      "logits/chosen": 0.0206298828125,
+      "logits/rejected": -0.099609375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -322.0,
+      "loss": 0.4253,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -4.0625,
+      "step": 585
+    },
+    {
+      "epoch": 0.7190184049079754,
+      "grad_norm": 36.65848945690881,
+      "learning_rate": 6.145454545454545e-07,
+      "logits/chosen": -0.1064453125,
+      "logits/rejected": -0.2158203125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3226,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.5625,
+      "step": 586
+    },
+    {
+      "epoch": 0.7202453987730061,
+      "grad_norm": 34.66555869797275,
+      "learning_rate": 6.142148760330578e-07,
+      "logits/chosen": -0.1396484375,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -332.0,
+      "logps/rejected": -308.0,
+      "loss": 0.319,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.5,
+      "step": 587
+    },
+    {
+      "epoch": 0.7214723926380369,
+      "grad_norm": 25.81313251350675,
+      "learning_rate": 6.138842975206612e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2328,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.9765625,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -4.34375,
+      "step": 588
+    },
+    {
+      "epoch": 0.7226993865030675,
+      "grad_norm": 39.45475953492192,
+      "learning_rate": 6.135537190082645e-07,
+      "logits/chosen": -0.2216796875,
+      "logits/rejected": -0.322265625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3428,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.7734375,
+      "rewards/margins": 4.84375,
+      "rewards/rejected": -4.0625,
+      "step": 589
+    },
+    {
+      "epoch": 0.7239263803680982,
+      "grad_norm": 28.739570926494295,
+      "learning_rate": 6.132231404958677e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.2392578125,
+      "logps/chosen": -364.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2629,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.53125,
+      "rewards/rejected": -4.375,
+      "step": 590
+    },
+    {
+      "epoch": 0.7251533742331289,
+      "grad_norm": 39.542558184509424,
+      "learning_rate": 6.12892561983471e-07,
+      "logits/chosen": -0.28125,
+      "logits/rejected": -0.474609375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2455,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -4.9375,
+      "step": 591
+    },
+    {
+      "epoch": 0.7263803680981595,
+      "grad_norm": 31.700565699717185,
+      "learning_rate": 6.125619834710743e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2372,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.6015625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.1875,
+      "step": 592
+    },
+    {
+      "epoch": 0.7276073619631902,
+      "grad_norm": 43.160388567583965,
+      "learning_rate": 6.122314049586776e-07,
+      "logits/chosen": -0.08251953125,
+      "logits/rejected": -0.341796875,
+      "logps/chosen": -442.0,
+      "logps/rejected": -366.0,
+      "loss": 0.3556,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.34375,
+      "rewards/rejected": -4.5625,
+      "step": 593
+    },
+    {
+      "epoch": 0.7288343558282209,
+      "grad_norm": 33.17336982955441,
+      "learning_rate": 6.11900826446281e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -356.0,
+      "logps/rejected": -314.0,
+      "loss": 0.2737,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.77734375,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.59375,
+      "step": 594
+    },
+    {
+      "epoch": 0.7300613496932515,
+      "grad_norm": 35.78921531146286,
+      "learning_rate": 6.115702479338843e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -444.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2188,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.40625,
+      "step": 595
+    },
+    {
+      "epoch": 0.7312883435582822,
+      "grad_norm": 39.132566907776244,
+      "learning_rate": 6.112396694214876e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3258,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.125,
+      "step": 596
+    },
+    {
+      "epoch": 0.7325153374233129,
+      "grad_norm": 36.16852227530472,
+      "learning_rate": 6.109090909090909e-07,
+      "logits/chosen": -0.03515625,
+      "logits/rejected": -0.2138671875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -290.0,
+      "loss": 0.3225,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.9375,
+      "step": 597
+    },
+    {
+      "epoch": 0.7337423312883435,
+      "grad_norm": 29.88526191086389,
+      "learning_rate": 6.105785123966942e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -438.0,
+      "logps/rejected": -318.0,
+      "loss": 0.204,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 7.21875,
+      "rewards/rejected": -5.125,
+      "step": 598
+    },
+    {
+      "epoch": 0.7349693251533742,
+      "grad_norm": 32.885582252909025,
+      "learning_rate": 6.102479338842974e-07,
+      "logits/chosen": -0.267578125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -448.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2847,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.6328125,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.53125,
+      "step": 599
+    },
+    {
+      "epoch": 0.7361963190184049,
+      "grad_norm": 37.722786017646094,
+      "learning_rate": 6.099173553719008e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -318.0,
+      "loss": 0.368,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.53125,
+      "step": 600
+    },
+    {
+      "epoch": 0.7374233128834355,
+      "grad_norm": 31.64488245915107,
+      "learning_rate": 6.095867768595041e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -306.0,
+      "loss": 0.2698,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.1826171875,
+      "rewards/margins": 6.125,
+      "rewards/rejected": -5.9375,
+      "step": 601
+    },
+    {
+      "epoch": 0.7386503067484662,
+      "grad_norm": 25.124092304435703,
+      "learning_rate": 6.092561983471074e-07,
+      "logits/chosen": -0.0076904296875,
+      "logits/rejected": -0.1708984375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -332.0,
+      "loss": 0.189,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.2421875,
+      "rewards/margins": 7.28125,
+      "rewards/rejected": -6.03125,
+      "step": 602
+    },
+    {
+      "epoch": 0.7398773006134969,
+      "grad_norm": 30.45546682074925,
+      "learning_rate": 6.089256198347107e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2864,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.546875,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.625,
+      "step": 603
+    },
+    {
+      "epoch": 0.7411042944785277,
+      "grad_norm": 31.94436328392251,
+      "learning_rate": 6.08595041322314e-07,
+      "logits/chosen": -0.2197265625,
+      "logits/rejected": -0.390625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2464,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.90625,
+      "step": 604
+    },
+    {
+      "epoch": 0.7423312883435583,
+      "grad_norm": 33.65745564504924,
+      "learning_rate": 6.082644628099173e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2598,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.984375,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -6.15625,
+      "step": 605
+    },
+    {
+      "epoch": 0.743558282208589,
+      "grad_norm": 38.63161562009038,
+      "learning_rate": 6.079338842975207e-07,
+      "logits/chosen": -0.2421875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3454,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1484375,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.3125,
+      "step": 606
+    },
+    {
+      "epoch": 0.7447852760736197,
+      "grad_norm": 33.230355625984764,
+      "learning_rate": 6.07603305785124e-07,
+      "logits/chosen": -0.12890625,
+      "logits/rejected": -0.349609375,
+      "logps/chosen": -396.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2611,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.48828125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -6.40625,
+      "step": 607
+    },
+    {
+      "epoch": 0.7460122699386503,
+      "grad_norm": 33.50647832359134,
+      "learning_rate": 6.072727272727272e-07,
+      "logits/chosen": -0.0216064453125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -396.0,
+      "logps/rejected": -342.0,
+      "loss": 0.2913,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.099609375,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -5.6875,
+      "step": 608
+    },
+    {
+      "epoch": 0.747239263803681,
+      "grad_norm": 35.459553964773974,
+      "learning_rate": 6.069421487603305e-07,
+      "logits/chosen": -0.1435546875,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2856,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.54296875,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -6.21875,
+      "step": 609
+    },
+    {
+      "epoch": 0.7484662576687117,
+      "grad_norm": 31.662560042647854,
+      "learning_rate": 6.066115702479338e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.419921875,
+      "logps/chosen": -444.0,
+      "logps/rejected": -376.0,
+      "loss": 0.2452,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.09375,
+      "rewards/margins": 7.46875,
+      "rewards/rejected": -6.375,
+      "step": 610
+    },
+    {
+      "epoch": 0.7496932515337423,
+      "grad_norm": 49.54193875289406,
+      "learning_rate": 6.062809917355371e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -432.0,
+      "logps/rejected": -362.0,
+      "loss": 0.3976,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": -0.1396484375,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -6.0,
+      "step": 611
+    },
+    {
+      "epoch": 0.750920245398773,
+      "grad_norm": 39.062720193674714,
+      "learning_rate": 6.059504132231405e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -392.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3777,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.546875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.40625,
+      "step": 612
+    },
+    {
+      "epoch": 0.7521472392638037,
+      "grad_norm": 35.9253243800041,
+      "learning_rate": 6.056198347107438e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.2216796875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3394,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.10107421875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.96875,
+      "step": 613
+    },
+    {
+      "epoch": 0.7533742331288343,
+      "grad_norm": 54.139929990418025,
+      "learning_rate": 6.052892561983471e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -420.0,
+      "logps/rejected": -344.0,
+      "loss": 0.3142,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.609375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.78125,
+      "step": 614
+    },
+    {
+      "epoch": 0.754601226993865,
+      "grad_norm": 37.247633297004576,
+      "learning_rate": 6.049586776859504e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -368.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3089,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.84765625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -6.0,
+      "step": 615
+    },
+    {
+      "epoch": 0.7558282208588957,
+      "grad_norm": 31.187596328354086,
+      "learning_rate": 6.046280991735537e-07,
+      "logits/chosen": -0.146484375,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -294.0,
+      "loss": 0.2528,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.10302734375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -5.46875,
+      "step": 616
+    },
+    {
+      "epoch": 0.7570552147239263,
+      "grad_norm": 39.842219715736135,
+      "learning_rate": 6.04297520661157e-07,
+      "logits/chosen": -0.125,
+      "logits/rejected": -0.208984375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3542,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.48828125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.3125,
+      "step": 617
+    },
+    {
+      "epoch": 0.758282208588957,
+      "grad_norm": 43.96609425022703,
+      "learning_rate": 6.039669421487603e-07,
+      "logits/chosen": -0.052734375,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3723,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.51953125,
+      "rewards/margins": 5.46875,
+      "rewards/rejected": -4.9375,
+      "step": 618
+    },
+    {
+      "epoch": 0.7595092024539877,
+      "grad_norm": 32.587670722826026,
+      "learning_rate": 6.036363636363636e-07,
+      "logits/chosen": -0.0791015625,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -426.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2331,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.90625,
+      "step": 619
+    },
+    {
+      "epoch": 0.7607361963190185,
+      "grad_norm": 44.41642388280846,
+      "learning_rate": 6.033057851239669e-07,
+      "logits/chosen": -0.0830078125,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3088,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.5234375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -5.21875,
+      "step": 620
+    },
+    {
+      "epoch": 0.7619631901840491,
+      "grad_norm": 34.02108964551065,
+      "learning_rate": 6.029752066115702e-07,
+      "logits/chosen": -0.19921875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -478.0,
+      "logps/rejected": -368.0,
+      "loss": 0.2661,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.171875,
+      "rewards/margins": 7.3125,
+      "rewards/rejected": -6.125,
+      "step": 621
+    },
+    {
+      "epoch": 0.7631901840490798,
+      "grad_norm": 31.832438088278238,
+      "learning_rate": 6.026446280991735e-07,
+      "logits/chosen": -0.0888671875,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -342.0,
+      "loss": 0.236,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.72265625,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -6.03125,
+      "step": 622
+    },
+    {
+      "epoch": 0.7644171779141105,
+      "grad_norm": 38.89511615880052,
+      "learning_rate": 6.023140495867769e-07,
+      "logits/chosen": -0.1181640625,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -452.0,
+      "logps/rejected": -376.0,
+      "loss": 0.3167,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.4453125,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.46875,
+      "step": 623
+    },
+    {
+      "epoch": 0.7656441717791411,
+      "grad_norm": 32.19915482428781,
+      "learning_rate": 6.019834710743802e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.357421875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2307,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.234375,
+      "rewards/margins": 7.78125,
+      "rewards/rejected": -6.53125,
+      "step": 624
+    },
+    {
+      "epoch": 0.7668711656441718,
+      "grad_norm": 32.198949298871604,
+      "learning_rate": 6.016528925619835e-07,
+      "logits/chosen": -0.1494140625,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2191,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.375,
+      "step": 625
+    },
+    {
+      "epoch": 0.7680981595092025,
+      "grad_norm": 40.391116665536885,
+      "learning_rate": 6.013223140495867e-07,
+      "logits/chosen": -0.265625,
+      "logits/rejected": -0.416015625,
+      "logps/chosen": -412.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3346,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.03125,
+      "step": 626
+    },
+    {
+      "epoch": 0.7693251533742331,
+      "grad_norm": 33.50213686948866,
+      "learning_rate": 6.0099173553719e-07,
+      "logits/chosen": -0.06396484375,
+      "logits/rejected": -0.154296875,
+      "logps/chosen": -310.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3297,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": -0.1552734375,
+      "rewards/margins": 5.0625,
+      "rewards/rejected": -5.21875,
+      "step": 627
+    },
+    {
+      "epoch": 0.7705521472392638,
+      "grad_norm": 37.25258333430096,
+      "learning_rate": 6.006611570247933e-07,
+      "logits/chosen": -0.0947265625,
+      "logits/rejected": -0.2392578125,
+      "logps/chosen": -376.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2886,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.046875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -4.90625,
+      "step": 628
+    },
+    {
+      "epoch": 0.7717791411042945,
+      "grad_norm": 23.68892864954692,
+      "learning_rate": 6.003305785123967e-07,
+      "logits/chosen": -0.1220703125,
+      "logits/rejected": -0.2470703125,
+      "logps/chosen": -348.0,
+      "logps/rejected": -284.0,
+      "loss": 0.1789,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.375,
+      "step": 629
+    },
+    {
+      "epoch": 0.7730061349693251,
+      "grad_norm": 33.10208379218177,
+      "learning_rate": 6e-07,
+      "logits/chosen": -0.060546875,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -370.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2663,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.7421875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.0625,
+      "step": 630
+    },
+    {
+      "epoch": 0.7742331288343558,
+      "grad_norm": 36.528138511875646,
+      "learning_rate": 5.996694214876033e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3321,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.53125,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -4.875,
+      "step": 631
+    },
+    {
+      "epoch": 0.7754601226993865,
+      "grad_norm": 36.37506860284647,
+      "learning_rate": 5.993388429752066e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -400.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3056,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.671875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.34375,
+      "step": 632
+    },
+    {
+      "epoch": 0.7766871165644171,
+      "grad_norm": 40.1484232566858,
+      "learning_rate": 5.990082644628099e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.21484375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3418,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.84375,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -4.65625,
+      "step": 633
+    },
+    {
+      "epoch": 0.7779141104294478,
+      "grad_norm": 32.83617762178678,
+      "learning_rate": 5.986776859504132e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -414.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2202,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 2.734375,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -5.1875,
+      "step": 634
+    },
+    {
+      "epoch": 0.7791411042944786,
+      "grad_norm": 29.099912658258905,
+      "learning_rate": 5.983471074380165e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -358.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2119,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.875,
+      "rewards/rejected": -4.84375,
+      "step": 635
+    },
+    {
+      "epoch": 0.7803680981595092,
+      "grad_norm": 39.54581126371982,
+      "learning_rate": 5.980165289256198e-07,
+      "logits/chosen": -0.1884765625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -366.0,
+      "loss": 0.3099,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.8046875,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -4.90625,
+      "step": 636
+    },
+    {
+      "epoch": 0.7815950920245399,
+      "grad_norm": 45.109718146173684,
+      "learning_rate": 5.976859504132231e-07,
+      "logits/chosen": -0.03125,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -300.0,
+      "loss": 0.3522,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.96875,
+      "rewards/margins": 5.96875,
+      "rewards/rejected": -5.0,
+      "step": 637
+    },
+    {
+      "epoch": 0.7828220858895706,
+      "grad_norm": 29.717665569657886,
+      "learning_rate": 5.973553719008264e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2213,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.6875,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -4.9375,
+      "step": 638
+    },
+    {
+      "epoch": 0.7840490797546013,
+      "grad_norm": 38.75212680036792,
+      "learning_rate": 5.970247933884297e-07,
+      "logits/chosen": -0.048828125,
+      "logits/rejected": -0.2265625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3316,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.6484375,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.34375,
+      "step": 639
+    },
+    {
+      "epoch": 0.7852760736196319,
+      "grad_norm": 39.02043679792906,
+      "learning_rate": 5.96694214876033e-07,
+      "logits/chosen": -0.2021484375,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -440.0,
+      "logps/rejected": -382.0,
+      "loss": 0.2894,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.15625,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -4.71875,
+      "step": 640
+    },
+    {
+      "epoch": 0.7865030674846626,
+      "grad_norm": 44.494628054471185,
+      "learning_rate": 5.963636363636364e-07,
+      "logits/chosen": -0.2373046875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -348.0,
+      "loss": 0.4553,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 1.734375,
+      "rewards/margins": 5.75,
+      "rewards/rejected": -4.03125,
+      "step": 641
+    },
+    {
+      "epoch": 0.7877300613496933,
+      "grad_norm": 27.200186618621586,
+      "learning_rate": 5.960330578512397e-07,
+      "logits/chosen": -0.1416015625,
+      "logits/rejected": -0.353515625,
+      "logps/chosen": -346.0,
+      "logps/rejected": -290.0,
+      "loss": 0.2487,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.5390625,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.5,
+      "step": 642
+    },
+    {
+      "epoch": 0.7889570552147239,
+      "grad_norm": 36.983988342062,
+      "learning_rate": 5.95702479338843e-07,
+      "logits/chosen": -0.275390625,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -398.0,
+      "logps/rejected": -372.0,
+      "loss": 0.2631,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.7109375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -4.53125,
+      "step": 643
+    },
+    {
+      "epoch": 0.7901840490797546,
+      "grad_norm": 39.431613095622076,
+      "learning_rate": 5.953719008264462e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2856,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.09375,
+      "step": 644
+    },
+    {
+      "epoch": 0.7914110429447853,
+      "grad_norm": 33.34429063414653,
+      "learning_rate": 5.950413223140495e-07,
+      "logits/chosen": -0.140625,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -328.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2445,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.796875,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -4.9375,
+      "step": 645
+    },
+    {
+      "epoch": 0.7926380368098159,
+      "grad_norm": 34.57872089054598,
+      "learning_rate": 5.947107438016528e-07,
+      "logits/chosen": -0.0859375,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2881,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.75,
+      "step": 646
+    },
+    {
+      "epoch": 0.7938650306748466,
+      "grad_norm": 34.439204602754096,
+      "learning_rate": 5.943801652892562e-07,
+      "logits/chosen": -0.1484375,
+      "logits/rejected": -0.296875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2819,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -5.125,
+      "step": 647
+    },
+    {
+      "epoch": 0.7950920245398773,
+      "grad_norm": 41.50373217977521,
+      "learning_rate": 5.940495867768595e-07,
+      "logits/chosen": -0.203125,
+      "logits/rejected": -0.2119140625,
+      "logps/chosen": -324.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3682,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.859375,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.84375,
+      "step": 648
+    },
+    {
+      "epoch": 0.7963190184049079,
+      "grad_norm": 43.045318488956525,
+      "learning_rate": 5.937190082644628e-07,
+      "logits/chosen": -0.0654296875,
+      "logits/rejected": -0.1650390625,
+      "logps/chosen": -362.0,
+      "logps/rejected": -324.0,
+      "loss": 0.491,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.21875,
+      "step": 649
+    },
+    {
+      "epoch": 0.7975460122699386,
+      "grad_norm": 41.49765653672477,
+      "learning_rate": 5.933884297520661e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3786,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 1.8125,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.09375,
+      "step": 650
+    },
+    {
+      "epoch": 0.7987730061349694,
+      "grad_norm": 38.92619113627064,
+      "learning_rate": 5.930578512396694e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -368.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2788,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.78125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -5.5625,
+      "step": 651
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 40.314158512080596,
+      "learning_rate": 5.927272727272728e-07,
+      "logits/chosen": -0.158203125,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3015,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -5.09375,
+      "step": 652
+    },
+    {
+      "epoch": 0.8012269938650307,
+      "grad_norm": 41.91917001542575,
+      "learning_rate": 5.92396694214876e-07,
+      "logits/chosen": -0.103515625,
+      "logits/rejected": -0.283203125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2786,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.625,
+      "step": 653
+    },
+    {
+      "epoch": 0.8024539877300614,
+      "grad_norm": 32.50826629622854,
+      "learning_rate": 5.920661157024793e-07,
+      "logits/chosen": -0.083984375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2325,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.34375,
+      "step": 654
+    },
+    {
+      "epoch": 0.803680981595092,
+      "grad_norm": 36.66838919465664,
+      "learning_rate": 5.917355371900826e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -384.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2743,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.15625,
+      "step": 655
+    },
+    {
+      "epoch": 0.8049079754601227,
+      "grad_norm": 40.353892524687666,
+      "learning_rate": 5.914049586776859e-07,
+      "logits/chosen": -0.080078125,
+      "logits/rejected": -0.2373046875,
+      "logps/chosen": -368.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3542,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 6.09375,
+      "rewards/rejected": -5.25,
+      "step": 656
+    },
+    {
+      "epoch": 0.8061349693251534,
+      "grad_norm": 32.71653938190321,
+      "learning_rate": 5.910743801652892e-07,
+      "logits/chosen": -0.2236328125,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3238,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.1875,
+      "step": 657
+    },
+    {
+      "epoch": 0.807361963190184,
+      "grad_norm": 35.63490892300594,
+      "learning_rate": 5.907438016528926e-07,
+      "logits/chosen": -0.23046875,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -388.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2672,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.99609375,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.375,
+      "step": 658
+    },
+    {
+      "epoch": 0.8085889570552147,
+      "grad_norm": 36.34676950346043,
+      "learning_rate": 5.904132231404959e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -362.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3323,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.2177734375,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -5.6875,
+      "step": 659
+    },
+    {
+      "epoch": 0.8098159509202454,
+      "grad_norm": 31.25262331885713,
+      "learning_rate": 5.900826446280992e-07,
+      "logits/chosen": -0.1826171875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -400.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2661,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.4296875,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.46875,
+      "step": 660
+    },
+    {
+      "epoch": 0.811042944785276,
+      "grad_norm": 44.04724360514248,
+      "learning_rate": 5.897520661157025e-07,
+      "logits/chosen": -0.11181640625,
+      "logits/rejected": -0.2236328125,
+      "logps/chosen": -360.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3716,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.3125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.46875,
+      "step": 661
+    },
+    {
+      "epoch": 0.8122699386503067,
+      "grad_norm": 28.306859195207647,
+      "learning_rate": 5.894214876033057e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -362.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2117,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -5.71875,
+      "step": 662
+    },
+    {
+      "epoch": 0.8134969325153374,
+      "grad_norm": 28.755676702838585,
+      "learning_rate": 5.89090909090909e-07,
+      "logits/chosen": -0.06689453125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -354.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2615,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.3671875,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.09375,
+      "step": 663
+    },
+    {
+      "epoch": 0.8147239263803681,
+      "grad_norm": 30.20251070806522,
+      "learning_rate": 5.887603305785124e-07,
+      "logits/chosen": -0.349609375,
+      "logits/rejected": -0.494140625,
+      "logps/chosen": -384.0,
+      "logps/rejected": -358.0,
+      "loss": 0.1821,
+      "rewards/accuracies": 0.9140625,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 7.84375,
+      "rewards/rejected": -6.625,
+      "step": 664
+    },
+    {
+      "epoch": 0.8159509202453987,
+      "grad_norm": 43.59493072844384,
+      "learning_rate": 5.884297520661157e-07,
+      "logits/chosen": -0.05029296875,
+      "logits/rejected": -0.2197265625,
+      "logps/chosen": -334.0,
+      "logps/rejected": -292.0,
+      "loss": 0.2979,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.5234375,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.78125,
+      "step": 665
+    },
+    {
+      "epoch": 0.8171779141104294,
+      "grad_norm": 42.64325196007558,
+      "learning_rate": 5.88099173553719e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.271484375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -348.0,
+      "loss": 0.368,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.21875,
+      "step": 666
+    },
+    {
+      "epoch": 0.8184049079754602,
+      "grad_norm": 34.27649786728366,
+      "learning_rate": 5.877685950413223e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.359375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2601,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.0625,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.75,
+      "step": 667
+    },
+    {
+      "epoch": 0.8196319018404908,
+      "grad_norm": 33.27456675278737,
+      "learning_rate": 5.874380165289256e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2841,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.15625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.40625,
+      "step": 668
+    },
+    {
+      "epoch": 0.8208588957055215,
+      "grad_norm": 41.4043511264187,
+      "learning_rate": 5.871074380165289e-07,
+      "logits/chosen": -0.212890625,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3017,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.375,
+      "step": 669
+    },
+    {
+      "epoch": 0.8220858895705522,
+      "grad_norm": 31.799223551777082,
+      "learning_rate": 5.867768595041323e-07,
+      "logits/chosen": -0.0191650390625,
+      "logits/rejected": -0.1650390625,
+      "logps/chosen": -324.0,
+      "logps/rejected": -296.0,
+      "loss": 0.2541,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.125,
+      "step": 670
+    },
+    {
+      "epoch": 0.8233128834355828,
+      "grad_norm": 31.913109572694793,
+      "learning_rate": 5.864462809917355e-07,
+      "logits/chosen": -0.1298828125,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2407,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.7890625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.0625,
+      "step": 671
+    },
+    {
+      "epoch": 0.8245398773006135,
+      "grad_norm": 32.62700830171627,
+      "learning_rate": 5.861157024793388e-07,
+      "logits/chosen": -0.109375,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -444.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2681,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.078125,
+      "rewards/margins": 7.25,
+      "rewards/rejected": -5.15625,
+      "step": 672
+    },
+    {
+      "epoch": 0.8257668711656442,
+      "grad_norm": 36.66995682138012,
+      "learning_rate": 5.857851239669421e-07,
+      "logits/chosen": -0.0498046875,
+      "logits/rejected": -0.28125,
+      "logps/chosen": -388.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2592,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.265625,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -4.625,
+      "step": 673
+    },
+    {
+      "epoch": 0.8269938650306748,
+      "grad_norm": 34.606200421728126,
+      "learning_rate": 5.854545454545454e-07,
+      "logits/chosen": -0.01300048828125,
+      "logits/rejected": -0.177734375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -324.0,
+      "loss": 0.2822,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.3359375,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -5.09375,
+      "step": 674
+    },
+    {
+      "epoch": 0.8282208588957055,
+      "grad_norm": 35.038586845163366,
+      "learning_rate": 5.851239669421487e-07,
+      "logits/chosen": -0.08642578125,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -424.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3168,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.28125,
+      "step": 675
+    },
+    {
+      "epoch": 0.8294478527607362,
+      "grad_norm": 35.465018732105385,
+      "learning_rate": 5.847933884297521e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -390.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2534,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.71875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.53125,
+      "step": 676
+    },
+    {
+      "epoch": 0.8306748466257668,
+      "grad_norm": 36.537944400776084,
+      "learning_rate": 5.844628099173554e-07,
+      "logits/chosen": -0.06982421875,
+      "logits/rejected": -0.2451171875,
+      "logps/chosen": -334.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2916,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.52734375,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.84375,
+      "step": 677
+    },
+    {
+      "epoch": 0.8319018404907975,
+      "grad_norm": 33.168213995988694,
+      "learning_rate": 5.841322314049587e-07,
+      "logits/chosen": -0.1005859375,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -418.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2768,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.95703125,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.25,
+      "step": 678
+    },
+    {
+      "epoch": 0.8331288343558282,
+      "grad_norm": 37.74716831610831,
+      "learning_rate": 5.838016528925619e-07,
+      "logits/chosen": -0.01220703125,
+      "logits/rejected": -0.1259765625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -314.0,
+      "loss": 0.3364,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.015625,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.8125,
+      "step": 679
+    },
+    {
+      "epoch": 0.8343558282208589,
+      "grad_norm": 34.79631984317766,
+      "learning_rate": 5.834710743801652e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -454.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2698,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.4375,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -5.65625,
+      "step": 680
+    },
+    {
+      "epoch": 0.8355828220858895,
+      "grad_norm": 27.30848566312445,
+      "learning_rate": 5.831404958677685e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -344.0,
+      "logps/rejected": -324.0,
+      "loss": 0.1854,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.0859375,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -6.25,
+      "step": 681
+    },
+    {
+      "epoch": 0.8368098159509203,
+      "grad_norm": 36.85289178579729,
+      "learning_rate": 5.828099173553719e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3333,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.81640625,
+      "rewards/margins": 6.46875,
+      "rewards/rejected": -5.65625,
+      "step": 682
+    },
+    {
+      "epoch": 0.838036809815951,
+      "grad_norm": 33.68839599065418,
+      "learning_rate": 5.824793388429752e-07,
+      "logits/chosen": -0.18359375,
+      "logits/rejected": -0.302734375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2767,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.203125,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.8125,
+      "step": 683
+    },
+    {
+      "epoch": 0.8392638036809816,
+      "grad_norm": 35.217179760941754,
+      "learning_rate": 5.821487603305785e-07,
+      "logits/chosen": 0.000904083251953125,
+      "logits/rejected": -0.1181640625,
+      "logps/chosen": -314.0,
+      "logps/rejected": -338.0,
+      "loss": 0.3304,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": -0.0341796875,
+      "rewards/margins": 5.40625,
+      "rewards/rejected": -5.4375,
+      "step": 684
+    },
+    {
+      "epoch": 0.8404907975460123,
+      "grad_norm": 34.3907812005455,
+      "learning_rate": 5.818181818181818e-07,
+      "logits/chosen": -0.11865234375,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -398.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2858,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.9375,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -6.59375,
+      "step": 685
+    },
+    {
+      "epoch": 0.841717791411043,
+      "grad_norm": 29.480838482306872,
+      "learning_rate": 5.814876033057851e-07,
+      "logits/chosen": -0.1328125,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2598,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.03125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.8125,
+      "step": 686
+    },
+    {
+      "epoch": 0.8429447852760736,
+      "grad_norm": 34.28684664510481,
+      "learning_rate": 5.811570247933884e-07,
+      "logits/chosen": -0.11767578125,
+      "logits/rejected": -0.263671875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2709,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.65625,
+      "step": 687
+    },
+    {
+      "epoch": 0.8441717791411043,
+      "grad_norm": 35.903679070126266,
+      "learning_rate": 5.808264462809917e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.326171875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2748,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.07958984375,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -6.125,
+      "step": 688
+    },
+    {
+      "epoch": 0.845398773006135,
+      "grad_norm": 37.11450186775036,
+      "learning_rate": 5.80495867768595e-07,
+      "logits/chosen": -0.12353515625,
+      "logits/rejected": -0.2197265625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3107,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.2109375,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.375,
+      "step": 689
+    },
+    {
+      "epoch": 0.8466257668711656,
+      "grad_norm": 49.86231404884124,
+      "learning_rate": 5.801652892561983e-07,
+      "logits/chosen": 0.004058837890625,
+      "logits/rejected": -0.1943359375,
+      "logps/chosen": -402.0,
+      "logps/rejected": -350.0,
+      "loss": 0.4457,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 6.8125,
+      "rewards/rejected": -5.3125,
+      "step": 690
+    },
+    {
+      "epoch": 0.8478527607361963,
+      "grad_norm": 30.320356705626047,
+      "learning_rate": 5.798347107438016e-07,
+      "logits/chosen": -0.298828125,
+      "logits/rejected": -0.41796875,
+      "logps/chosen": -356.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2435,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.71875,
+      "step": 691
+    },
+    {
+      "epoch": 0.849079754601227,
+      "grad_norm": 43.95989415122109,
+      "learning_rate": 5.795041322314049e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.29296875,
+      "logps/chosen": -408.0,
+      "logps/rejected": -336.0,
+      "loss": 0.282,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.703125,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -5.34375,
+      "step": 692
+    },
+    {
+      "epoch": 0.8503067484662576,
+      "grad_norm": 33.0652784939274,
+      "learning_rate": 5.791735537190082e-07,
+      "logits/chosen": -0.1474609375,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -442.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2934,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.203125,
+      "rewards/margins": 7.8125,
+      "rewards/rejected": -5.625,
+      "step": 693
+    },
+    {
+      "epoch": 0.8515337423312883,
+      "grad_norm": 46.044690743816275,
+      "learning_rate": 5.788429752066116e-07,
+      "logits/chosen": -0.13671875,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -370.0,
+      "logps/rejected": -330.0,
+      "loss": 0.4005,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.34375,
+      "rewards/rejected": -4.5625,
+      "step": 694
+    },
+    {
+      "epoch": 0.852760736196319,
+      "grad_norm": 47.292538608776496,
+      "learning_rate": 5.785123966942149e-07,
+      "logits/chosen": -0.11376953125,
+      "logits/rejected": -0.208984375,
+      "logps/chosen": -328.0,
+      "logps/rejected": -302.0,
+      "loss": 0.4681,
+      "rewards/accuracies": 0.7578125,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 4.625,
+      "rewards/rejected": -3.90625,
+      "step": 695
+    },
+    {
+      "epoch": 0.8539877300613496,
+      "grad_norm": 38.43478974924762,
+      "learning_rate": 5.781818181818182e-07,
+      "logits/chosen": -0.07568359375,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -316.0,
+      "loss": 0.3104,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.5625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -4.6875,
+      "step": 696
+    },
+    {
+      "epoch": 0.8552147239263803,
+      "grad_norm": 58.82691398412282,
+      "learning_rate": 5.778512396694214e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.2734375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3128,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 2.140625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -4.78125,
+      "step": 697
+    },
+    {
+      "epoch": 0.8564417177914111,
+      "grad_norm": 37.608292873182215,
+      "learning_rate": 5.775206611570247e-07,
+      "logits/chosen": -0.080078125,
+      "logits/rejected": -0.216796875,
+      "logps/chosen": -392.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2946,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.6640625,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -4.84375,
+      "step": 698
+    },
+    {
+      "epoch": 0.8576687116564418,
+      "grad_norm": 41.848029972370156,
+      "learning_rate": 5.77190082644628e-07,
+      "logits/chosen": -0.1455078125,
+      "logits/rejected": -0.2421875,
+      "logps/chosen": -334.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2843,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.91015625,
+      "rewards/margins": 5.625,
+      "rewards/rejected": -4.71875,
+      "step": 699
+    },
+    {
+      "epoch": 0.8588957055214724,
+      "grad_norm": 39.920137771888676,
+      "learning_rate": 5.768595041322314e-07,
+      "logits/chosen": -0.054931640625,
+      "logits/rejected": -0.1640625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -320.0,
+      "loss": 0.3259,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.34375,
+      "rewards/margins": 5.59375,
+      "rewards/rejected": -4.25,
+      "step": 700
+    },
+    {
+      "epoch": 0.8601226993865031,
+      "grad_norm": 41.52681124295287,
+      "learning_rate": 5.765289256198347e-07,
+      "logits/chosen": -0.025146484375,
+      "logits/rejected": -0.138671875,
+      "logps/chosen": -376.0,
+      "logps/rejected": -322.0,
+      "loss": 0.385,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 5.84375,
+      "rewards/rejected": -4.78125,
+      "step": 701
+    },
+    {
+      "epoch": 0.8613496932515338,
+      "grad_norm": 28.252991079387233,
+      "learning_rate": 5.76198347107438e-07,
+      "logits/chosen": -0.166015625,
+      "logits/rejected": -0.458984375,
+      "logps/chosen": -394.0,
+      "logps/rejected": -330.0,
+      "loss": 0.1786,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 2.125,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -5.21875,
+      "step": 702
+    },
+    {
+      "epoch": 0.8625766871165644,
+      "grad_norm": 31.56795802217109,
+      "learning_rate": 5.758677685950413e-07,
+      "logits/chosen": -0.046142578125,
+      "logits/rejected": -0.1884765625,
+      "logps/chosen": -410.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2909,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -4.3125,
+      "step": 703
+    },
+    {
+      "epoch": 0.8638036809815951,
+      "grad_norm": 34.11635728276886,
+      "learning_rate": 5.755371900826446e-07,
+      "logits/chosen": -0.1201171875,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -320.0,
+      "logps/rejected": -298.0,
+      "loss": 0.3105,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 6.1875,
+      "rewards/rejected": -4.9375,
+      "step": 704
+    },
+    {
+      "epoch": 0.8650306748466258,
+      "grad_norm": 38.95077133588588,
+      "learning_rate": 5.75206611570248e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.365234375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2767,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 2.265625,
+      "rewards/margins": 7.15625,
+      "rewards/rejected": -4.875,
+      "step": 705
+    },
+    {
+      "epoch": 0.8662576687116564,
+      "grad_norm": 35.267737415428456,
+      "learning_rate": 5.748760330578512e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -294.0,
+      "loss": 0.3068,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.4609375,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.53125,
+      "step": 706
+    },
+    {
+      "epoch": 0.8674846625766871,
+      "grad_norm": 34.32550780952695,
+      "learning_rate": 5.745454545454545e-07,
+      "logits/chosen": -0.1845703125,
+      "logits/rejected": -0.34765625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -304.0,
+      "loss": 0.2683,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.1796875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.25,
+      "step": 707
+    },
+    {
+      "epoch": 0.8687116564417178,
+      "grad_norm": 31.650191661242673,
+      "learning_rate": 5.742148760330578e-07,
+      "logits/chosen": -0.1669921875,
+      "logits/rejected": -0.384765625,
+      "logps/chosen": -370.0,
+      "logps/rejected": -312.0,
+      "loss": 0.2379,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.6796875,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -4.875,
+      "step": 708
+    },
+    {
+      "epoch": 0.8699386503067484,
+      "grad_norm": 25.32301991794396,
+      "learning_rate": 5.738842975206611e-07,
+      "logits/chosen": -0.06640625,
+      "logits/rejected": -0.23046875,
+      "logps/chosen": -434.0,
+      "logps/rejected": -340.0,
+      "loss": 0.185,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 1.9765625,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -5.53125,
+      "step": 709
+    },
+    {
+      "epoch": 0.8711656441717791,
+      "grad_norm": 33.22472192244502,
+      "learning_rate": 5.735537190082644e-07,
+      "logits/chosen": -0.017333984375,
+      "logits/rejected": -0.15234375,
+      "logps/chosen": -362.0,
+      "logps/rejected": -312.0,
+      "loss": 0.299,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1015625,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.625,
+      "step": 710
+    },
+    {
+      "epoch": 0.8723926380368098,
+      "grad_norm": 29.02267964751041,
+      "learning_rate": 5.732231404958678e-07,
+      "logits/chosen": -0.042724609375,
+      "logits/rejected": -0.236328125,
+      "logps/chosen": -382.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2551,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5078125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -4.78125,
+      "step": 711
+    },
+    {
+      "epoch": 0.8736196319018404,
+      "grad_norm": 40.668984053435686,
+      "learning_rate": 5.728925619834711e-07,
+      "logits/chosen": -0.03857421875,
+      "logits/rejected": -0.1923828125,
+      "logps/chosen": -404.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3587,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.9921875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -4.375,
+      "step": 712
+    },
+    {
+      "epoch": 0.8748466257668711,
+      "grad_norm": 33.08084767770349,
+      "learning_rate": 5.725619834710744e-07,
+      "logits/chosen": -0.1279296875,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -320.0,
+      "loss": 0.2665,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.28125,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.15625,
+      "step": 713
+    },
+    {
+      "epoch": 0.8760736196319019,
+      "grad_norm": 36.45936464644564,
+      "learning_rate": 5.722314049586777e-07,
+      "logits/chosen": -0.0159912109375,
+      "logits/rejected": -0.2333984375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2938,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -4.75,
+      "step": 714
+    },
+    {
+      "epoch": 0.8773006134969326,
+      "grad_norm": 30.913206903477224,
+      "learning_rate": 5.719008264462809e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.310546875,
+      "logps/chosen": -382.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2684,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.3515625,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.53125,
+      "step": 715
+    },
+    {
+      "epoch": 0.8785276073619632,
+      "grad_norm": 41.616080953336535,
+      "learning_rate": 5.715702479338842e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -380.0,
+      "logps/rejected": -312.0,
+      "loss": 0.3142,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.421875,
+      "rewards/margins": 6.03125,
+      "rewards/rejected": -4.625,
+      "step": 716
+    },
+    {
+      "epoch": 0.8797546012269939,
+      "grad_norm": 34.18654539125182,
+      "learning_rate": 5.712396694214876e-07,
+      "logits/chosen": -0.09423828125,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -276.0,
+      "loss": 0.2911,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.9296875,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.3125,
+      "step": 717
+    },
+    {
+      "epoch": 0.8809815950920246,
+      "grad_norm": 28.09115304120503,
+      "learning_rate": 5.709090909090909e-07,
+      "logits/chosen": -0.1875,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -398.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2254,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.859375,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.0625,
+      "step": 718
+    },
+    {
+      "epoch": 0.8822085889570552,
+      "grad_norm": 41.08426714668054,
+      "learning_rate": 5.705785123966942e-07,
+      "logits/chosen": -0.2275390625,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -406.0,
+      "logps/rejected": -354.0,
+      "loss": 0.3011,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 2.25,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -4.9375,
+      "step": 719
+    },
+    {
+      "epoch": 0.8834355828220859,
+      "grad_norm": 28.518925312083375,
+      "learning_rate": 5.702479338842975e-07,
+      "logits/chosen": -0.22265625,
+      "logits/rejected": -0.43359375,
+      "logps/chosen": -410.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2223,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.265625,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -4.78125,
+      "step": 720
+    },
+    {
+      "epoch": 0.8846625766871166,
+      "grad_norm": 26.149570167864507,
+      "learning_rate": 5.699173553719008e-07,
+      "logits/chosen": -0.20703125,
+      "logits/rejected": -0.46875,
+      "logps/chosen": -410.0,
+      "logps/rejected": -352.0,
+      "loss": 0.2073,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.890625,
+      "rewards/margins": 7.96875,
+      "rewards/rejected": -6.09375,
+      "step": 721
+    },
+    {
+      "epoch": 0.8858895705521472,
+      "grad_norm": 30.929567326750046,
+      "learning_rate": 5.695867768595041e-07,
+      "logits/chosen": -0.1923828125,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2229,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.859375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.125,
+      "step": 722
+    },
+    {
+      "epoch": 0.8871165644171779,
+      "grad_norm": 34.37293415398814,
+      "learning_rate": 5.692561983471075e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -372.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2808,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.3125,
+      "rewards/margins": 6.75,
+      "rewards/rejected": -5.4375,
+      "step": 723
+    },
+    {
+      "epoch": 0.8883435582822086,
+      "grad_norm": 33.90677929130864,
+      "learning_rate": 5.689256198347107e-07,
+      "logits/chosen": -0.1806640625,
+      "logits/rejected": -0.40625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -308.0,
+      "loss": 0.254,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.7265625,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.21875,
+      "step": 724
+    },
+    {
+      "epoch": 0.8895705521472392,
+      "grad_norm": 34.97307578381243,
+      "learning_rate": 5.68595041322314e-07,
+      "logits/chosen": -0.055419921875,
+      "logits/rejected": -0.2314453125,
+      "logps/chosen": -400.0,
+      "logps/rejected": -326.0,
+      "loss": 0.3044,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.453125,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.15625,
+      "step": 725
+    },
+    {
+      "epoch": 0.8907975460122699,
+      "grad_norm": 28.627835648582934,
+      "learning_rate": 5.682644628099173e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -336.0,
+      "logps/rejected": -288.0,
+      "loss": 0.2135,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.9765625,
+      "rewards/margins": 6.5625,
+      "rewards/rejected": -5.5625,
+      "step": 726
+    },
+    {
+      "epoch": 0.8920245398773006,
+      "grad_norm": 34.81272514902735,
+      "learning_rate": 5.679338842975206e-07,
+      "logits/chosen": -0.15625,
+      "logits/rejected": -0.306640625,
+      "logps/chosen": -368.0,
+      "logps/rejected": -324.0,
+      "loss": 0.248,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.5234375,
+      "rewards/margins": 6.6875,
+      "rewards/rejected": -5.15625,
+      "step": 727
+    },
+    {
+      "epoch": 0.8932515337423312,
+      "grad_norm": 37.8936877345969,
+      "learning_rate": 5.676033057851239e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -398.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2946,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 2.015625,
+      "rewards/margins": 7.65625,
+      "rewards/rejected": -5.625,
+      "step": 728
+    },
+    {
+      "epoch": 0.894478527607362,
+      "grad_norm": 29.90567314985844,
+      "learning_rate": 5.672727272727273e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.333984375,
+      "logps/chosen": -390.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2088,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 7.6875,
+      "rewards/rejected": -6.0625,
+      "step": 729
+    },
+    {
+      "epoch": 0.8957055214723927,
+      "grad_norm": 31.408249750098918,
+      "learning_rate": 5.669421487603306e-07,
+      "logits/chosen": -0.2578125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -398.0,
+      "logps/rejected": -364.0,
+      "loss": 0.2204,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 2.1875,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -5.75,
+      "step": 730
+    },
+    {
+      "epoch": 0.8969325153374234,
+      "grad_norm": 33.91624972638253,
+      "learning_rate": 5.666115702479339e-07,
+      "logits/chosen": -0.052734375,
+      "logits/rejected": -0.1845703125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2736,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 7.1875,
+      "rewards/rejected": -5.8125,
+      "step": 731
+    },
+    {
+      "epoch": 0.898159509202454,
+      "grad_norm": 40.7773464371931,
+      "learning_rate": 5.662809917355372e-07,
+      "logits/chosen": -0.2080078125,
+      "logits/rejected": -0.376953125,
+      "logps/chosen": -372.0,
+      "logps/rejected": -356.0,
+      "loss": 0.3647,
+      "rewards/accuracies": 0.78125,
+      "rewards/chosen": 1.390625,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.375,
+      "step": 732
+    },
+    {
+      "epoch": 0.8993865030674847,
+      "grad_norm": 31.001224216873553,
+      "learning_rate": 5.659504132231404e-07,
+      "logits/chosen": -0.10791015625,
+      "logits/rejected": -0.267578125,
+      "logps/chosen": -356.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2212,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.0546875,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.84375,
+      "step": 733
+    },
+    {
+      "epoch": 0.9006134969325154,
+      "grad_norm": 37.803749225628884,
+      "learning_rate": 5.656198347107437e-07,
+      "logits/chosen": -0.08837890625,
+      "logits/rejected": -0.275390625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2954,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.2578125,
+      "rewards/margins": 7.5,
+      "rewards/rejected": -6.25,
+      "step": 734
+    },
+    {
+      "epoch": 0.901840490797546,
+      "grad_norm": 30.18047116909555,
+      "learning_rate": 5.652892561983471e-07,
+      "logits/chosen": -0.1943359375,
+      "logits/rejected": -0.427734375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -372.0,
+      "loss": 0.1824,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 8.125,
+      "rewards/rejected": -6.8125,
+      "step": 735
+    },
+    {
+      "epoch": 0.9030674846625767,
+      "grad_norm": 38.572832638835585,
+      "learning_rate": 5.649586776859504e-07,
+      "logits/chosen": -0.025634765625,
+      "logits/rejected": -0.1552734375,
+      "logps/chosen": -356.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3608,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.0908203125,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -6.3125,
+      "step": 736
+    },
+    {
+      "epoch": 0.9042944785276074,
+      "grad_norm": 46.258369054797996,
+      "learning_rate": 5.646280991735537e-07,
+      "logits/chosen": -0.12451171875,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -334.0,
+      "loss": 0.4383,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.1826171875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -6.25,
+      "step": 737
+    },
+    {
+      "epoch": 0.905521472392638,
+      "grad_norm": 32.39664225178629,
+      "learning_rate": 5.64297520661157e-07,
+      "logits/chosen": -0.185546875,
+      "logits/rejected": -0.31640625,
+      "logps/chosen": -442.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2588,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.3046875,
+      "rewards/margins": 7.5625,
+      "rewards/rejected": -6.28125,
+      "step": 738
+    },
+    {
+      "epoch": 0.9067484662576687,
+      "grad_norm": 34.18126494383032,
+      "learning_rate": 5.639669421487603e-07,
+      "logits/chosen": -0.11328125,
+      "logits/rejected": -0.25390625,
+      "logps/chosen": -344.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2836,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.2470703125,
+      "rewards/margins": 6.40625,
+      "rewards/rejected": -6.15625,
+      "step": 739
+    },
+    {
+      "epoch": 0.9079754601226994,
+      "grad_norm": 35.15083279583394,
+      "learning_rate": 5.636363636363636e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.345703125,
+      "logps/chosen": -342.0,
+      "logps/rejected": -318.0,
+      "loss": 0.3072,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.2431640625,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -6.6875,
+      "step": 740
+    },
+    {
+      "epoch": 0.90920245398773,
+      "grad_norm": 33.535600643328806,
+      "learning_rate": 5.63305785123967e-07,
+      "logits/chosen": -0.16796875,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -354.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2851,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.125,
+      "rewards/margins": 7.9375,
+      "rewards/rejected": -6.8125,
+      "step": 741
+    },
+    {
+      "epoch": 0.9104294478527607,
+      "grad_norm": 31.52809779595991,
+      "learning_rate": 5.629752066115702e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.3671875,
+      "logps/chosen": -364.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2244,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.140625,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.90625,
+      "step": 742
+    },
+    {
+      "epoch": 0.9116564417177914,
+      "grad_norm": 35.37117075886455,
+      "learning_rate": 5.626446280991735e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.212890625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -380.0,
+      "loss": 0.2769,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.8125,
+      "step": 743
+    },
+    {
+      "epoch": 0.912883435582822,
+      "grad_norm": 26.42035749866701,
+      "learning_rate": 5.623140495867768e-07,
+      "logits/chosen": -0.1259765625,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -330.0,
+      "loss": 0.1917,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.328125,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.75,
+      "step": 744
+    },
+    {
+      "epoch": 0.9141104294478528,
+      "grad_norm": 35.853268232755404,
+      "learning_rate": 5.619834710743801e-07,
+      "logits/chosen": -0.08251953125,
+      "logits/rejected": -0.1865234375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3009,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.390625,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.875,
+      "step": 745
+    },
+    {
+      "epoch": 0.9153374233128835,
+      "grad_norm": 43.35686834196019,
+      "learning_rate": 5.616528925619834e-07,
+      "logits/chosen": -0.0849609375,
+      "logits/rejected": -0.255859375,
+      "logps/chosen": -404.0,
+      "logps/rejected": -346.0,
+      "loss": 0.317,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.86328125,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -6.0,
+      "step": 746
+    },
+    {
+      "epoch": 0.9165644171779141,
+      "grad_norm": 34.17444251605582,
+      "learning_rate": 5.613223140495868e-07,
+      "logits/chosen": -0.26171875,
+      "logits/rejected": -0.380859375,
+      "logps/chosen": -338.0,
+      "logps/rejected": -336.0,
+      "loss": 0.2811,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.41796875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.78125,
+      "step": 747
+    },
+    {
+      "epoch": 0.9177914110429448,
+      "grad_norm": 40.3104413419788,
+      "learning_rate": 5.609917355371901e-07,
+      "logits/chosen": -0.05224609375,
+      "logits/rejected": -0.169921875,
+      "logps/chosen": -374.0,
+      "logps/rejected": -342.0,
+      "loss": 0.3882,
+      "rewards/accuracies": 0.796875,
+      "rewards/chosen": 0.0279541015625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -5.625,
+      "step": 748
+    },
+    {
+      "epoch": 0.9190184049079755,
+      "grad_norm": 40.57709610969512,
+      "learning_rate": 5.606611570247934e-07,
+      "logits/chosen": -0.1865234375,
+      "logits/rejected": -0.396484375,
+      "logps/chosen": -416.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3642,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.98828125,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -6.09375,
+      "step": 749
+    },
+    {
+      "epoch": 0.9202453987730062,
+      "grad_norm": 43.682412563973024,
+      "learning_rate": 5.603305785123967e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.298828125,
+      "logps/chosen": -406.0,
+      "logps/rejected": -334.0,
+      "loss": 0.3704,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.63671875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.875,
+      "step": 750
+    },
+    {
+      "epoch": 0.9214723926380368,
+      "grad_norm": 42.01608666567515,
+      "learning_rate": 5.599999999999999e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.337890625,
+      "logps/chosen": -422.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3006,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 7.4375,
+      "rewards/rejected": -6.28125,
+      "step": 751
+    },
+    {
+      "epoch": 0.9226993865030675,
+      "grad_norm": 32.96496448782055,
+      "learning_rate": 5.596694214876033e-07,
+      "logits/chosen": -0.169921875,
+      "logits/rejected": -0.431640625,
+      "logps/chosen": -418.0,
+      "logps/rejected": -354.0,
+      "loss": 0.2435,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.3984375,
+      "rewards/margins": 7.75,
+      "rewards/rejected": -6.34375,
+      "step": 752
+    },
+    {
+      "epoch": 0.9239263803680982,
+      "grad_norm": 35.77444345136624,
+      "learning_rate": 5.593388429752066e-07,
+      "logits/chosen": -0.048095703125,
+      "logits/rejected": -0.1923828125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2851,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.796875,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.5625,
+      "step": 753
+    },
+    {
+      "epoch": 0.9251533742331288,
+      "grad_norm": 39.56695959824003,
+      "learning_rate": 5.590082644628099e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.287109375,
+      "logps/chosen": -386.0,
+      "logps/rejected": -332.0,
+      "loss": 0.3588,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.703125,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -5.59375,
+      "step": 754
+    },
+    {
+      "epoch": 0.9263803680981595,
+      "grad_norm": 33.47777157003214,
+      "learning_rate": 5.586776859504132e-07,
+      "logits/chosen": -0.162109375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -376.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2421,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.890625,
+      "rewards/margins": 6.59375,
+      "rewards/rejected": -5.6875,
+      "step": 755
+    },
+    {
+      "epoch": 0.9276073619631902,
+      "grad_norm": 30.73963254337421,
+      "learning_rate": 5.583471074380165e-07,
+      "logits/chosen": -0.0732421875,
+      "logits/rejected": -0.1669921875,
+      "logps/chosen": -332.0,
+      "logps/rejected": -308.0,
+      "loss": 0.248,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 0.322265625,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.75,
+      "step": 756
+    },
+    {
+      "epoch": 0.9288343558282208,
+      "grad_norm": 25.863947966401952,
+      "learning_rate": 5.580165289256198e-07,
+      "logits/chosen": -0.0252685546875,
+      "logits/rejected": -0.251953125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2117,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.640625,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -5.375,
+      "step": 757
+    },
+    {
+      "epoch": 0.9300613496932515,
+      "grad_norm": 40.231833773135115,
+      "learning_rate": 5.576859504132232e-07,
+      "logits/chosen": -0.08837890625,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -366.0,
+      "logps/rejected": -304.0,
+      "loss": 0.3497,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.46875,
+      "rewards/margins": 6.0625,
+      "rewards/rejected": -5.59375,
+      "step": 758
+    },
+    {
+      "epoch": 0.9312883435582822,
+      "grad_norm": 34.14796678847348,
+      "learning_rate": 5.573553719008265e-07,
+      "logits/chosen": -0.0595703125,
+      "logits/rejected": -0.291015625,
+      "logps/chosen": -386.0,
+      "logps/rejected": -326.0,
+      "loss": 0.2511,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 6.53125,
+      "rewards/rejected": -5.34375,
+      "step": 759
+    },
+    {
+      "epoch": 0.9325153374233128,
+      "grad_norm": 29.867916639384045,
+      "learning_rate": 5.570247933884297e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -432.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2724,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 2.109375,
+      "rewards/margins": 7.8125,
+      "rewards/rejected": -5.71875,
+      "step": 760
+    },
+    {
+      "epoch": 0.9337423312883436,
+      "grad_norm": 38.414301746929866,
+      "learning_rate": 5.56694214876033e-07,
+      "logits/chosen": -0.0264892578125,
+      "logits/rejected": -0.1669921875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -340.0,
+      "loss": 0.2874,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.375,
+      "step": 761
+    },
+    {
+      "epoch": 0.9349693251533743,
+      "grad_norm": 40.447076836962,
+      "learning_rate": 5.563636363636363e-07,
+      "logits/chosen": -0.01397705078125,
+      "logits/rejected": -0.19921875,
+      "logps/chosen": -420.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2888,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.6875,
+      "step": 762
+    },
+    {
+      "epoch": 0.9361963190184049,
+      "grad_norm": 31.879754151772115,
+      "learning_rate": 5.560330578512396e-07,
+      "logits/chosen": -0.07373046875,
+      "logits/rejected": -0.171875,
+      "logps/chosen": -390.0,
+      "logps/rejected": -346.0,
+      "loss": 0.2711,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.5,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.4375,
+      "step": 763
+    },
+    {
+      "epoch": 0.9374233128834356,
+      "grad_norm": 35.145542304507835,
+      "learning_rate": 5.55702479338843e-07,
+      "logits/chosen": -0.28515625,
+      "logits/rejected": -0.447265625,
+      "logps/chosen": -322.0,
+      "logps/rejected": -302.0,
+      "loss": 0.2896,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.26171875,
+      "rewards/margins": 6.28125,
+      "rewards/rejected": -6.0,
+      "step": 764
+    },
+    {
+      "epoch": 0.9386503067484663,
+      "grad_norm": 40.67019743057879,
+      "learning_rate": 5.553719008264463e-07,
+      "logits/chosen": -0.0595703125,
+      "logits/rejected": -0.19140625,
+      "logps/chosen": -394.0,
+      "logps/rejected": -306.0,
+      "loss": 0.3457,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.76171875,
+      "rewards/margins": 5.71875,
+      "rewards/rejected": -4.96875,
+      "step": 765
+    },
+    {
+      "epoch": 0.939877300613497,
+      "grad_norm": 32.24931969694431,
+      "learning_rate": 5.550413223140496e-07,
+      "logits/chosen": -0.11767578125,
+      "logits/rejected": -0.3203125,
+      "logps/chosen": -436.0,
+      "logps/rejected": -326.0,
+      "loss": 0.274,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.375,
+      "rewards/margins": 6.9375,
+      "rewards/rejected": -5.5625,
+      "step": 766
+    },
+    {
+      "epoch": 0.9411042944785276,
+      "grad_norm": 34.46831261260643,
+      "learning_rate": 5.547107438016529e-07,
+      "logits/chosen": -0.134765625,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -378.0,
+      "logps/rejected": -324.0,
+      "loss": 0.3035,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.953125,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.9375,
+      "step": 767
+    },
+    {
+      "epoch": 0.9423312883435583,
+      "grad_norm": 36.15353177119284,
+      "learning_rate": 5.543801652892561e-07,
+      "logits/chosen": -0.150390625,
+      "logits/rejected": -0.33203125,
+      "logps/chosen": -402.0,
+      "logps/rejected": -372.0,
+      "loss": 0.316,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.359375,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -5.75,
+      "step": 768
+    },
+    {
+      "epoch": 0.943558282208589,
+      "grad_norm": 29.41748843520645,
+      "learning_rate": 5.540495867768594e-07,
+      "logits/chosen": -0.1611328125,
+      "logits/rejected": -0.30859375,
+      "logps/chosen": -372.0,
+      "logps/rejected": -332.0,
+      "loss": 0.2432,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.6328125,
+      "rewards/margins": 7.03125,
+      "rewards/rejected": -6.375,
+      "step": 769
+    },
+    {
+      "epoch": 0.9447852760736196,
+      "grad_norm": 41.73128837710126,
+      "learning_rate": 5.537190082644628e-07,
+      "logits/chosen": -0.1748046875,
+      "logits/rejected": -0.42578125,
+      "logps/chosen": -408.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3004,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.6171875,
+      "rewards/margins": 7.09375,
+      "rewards/rejected": -5.46875,
+      "step": 770
+    },
+    {
+      "epoch": 0.9460122699386503,
+      "grad_norm": 42.981692808778455,
+      "learning_rate": 5.533884297520661e-07,
+      "logits/chosen": -0.1689453125,
+      "logits/rejected": -0.201171875,
+      "logps/chosen": -422.0,
+      "logps/rejected": -346.0,
+      "loss": 0.429,
+      "rewards/accuracies": 0.7734375,
+      "rewards/chosen": 0.0869140625,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -5.5625,
+      "step": 771
+    },
+    {
+      "epoch": 0.947239263803681,
+      "grad_norm": 35.00554577923317,
+      "learning_rate": 5.530578512396694e-07,
+      "logits/chosen": -0.038330078125,
+      "logits/rejected": -0.259765625,
+      "logps/chosen": -416.0,
+      "logps/rejected": -350.0,
+      "loss": 0.3114,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 1.1640625,
+      "rewards/margins": 6.84375,
+      "rewards/rejected": -5.6875,
+      "step": 772
+    },
+    {
+      "epoch": 0.9484662576687116,
+      "grad_norm": 36.88335000648106,
+      "learning_rate": 5.527272727272727e-07,
+      "logits/chosen": -0.0361328125,
+      "logits/rejected": -0.2255859375,
+      "logps/chosen": -366.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3175,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.1513671875,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.78125,
+      "step": 773
+    },
+    {
+      "epoch": 0.9496932515337423,
+      "grad_norm": 38.52772066032803,
+      "learning_rate": 5.52396694214876e-07,
+      "logits/chosen": -0.058837890625,
+      "logits/rejected": -0.2001953125,
+      "logps/chosen": -352.0,
+      "logps/rejected": -310.0,
+      "loss": 0.3665,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": -0.330078125,
+      "rewards/margins": 4.75,
+      "rewards/rejected": -5.0625,
+      "step": 774
+    },
+    {
+      "epoch": 0.950920245398773,
+      "grad_norm": 38.30272199053397,
+      "learning_rate": 5.520661157024793e-07,
+      "logits/chosen": 0.0306396484375,
+      "logits/rejected": -0.162109375,
+      "logps/chosen": -406.0,
+      "logps/rejected": -348.0,
+      "loss": 0.3245,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.03125,
+      "step": 775
+    },
+    {
+      "epoch": 0.9521472392638037,
+      "grad_norm": 33.16873912890908,
+      "learning_rate": 5.517355371900827e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.35546875,
+      "logps/chosen": -404.0,
+      "logps/rejected": -352.0,
+      "loss": 0.3113,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.1171875,
+      "rewards/margins": 6.5,
+      "rewards/rejected": -5.375,
+      "step": 776
+    },
+    {
+      "epoch": 0.9533742331288344,
+      "grad_norm": 35.432794054015964,
+      "learning_rate": 5.514049586776859e-07,
+      "logits/chosen": -0.142578125,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -396.0,
+      "logps/rejected": -330.0,
+      "loss": 0.3098,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 0.5625,
+      "rewards/margins": 6.375,
+      "rewards/rejected": -5.8125,
+      "step": 777
+    },
+    {
+      "epoch": 0.9546012269938651,
+      "grad_norm": 35.189286152943,
+      "learning_rate": 5.510743801652892e-07,
+      "logits/chosen": -0.1083984375,
+      "logits/rejected": -0.318359375,
+      "logps/chosen": -388.0,
+      "logps/rejected": -314.0,
+      "loss": 0.251,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.443359375,
+      "rewards/margins": 5.78125,
+      "rewards/rejected": -5.34375,
+      "step": 778
+    },
+    {
+      "epoch": 0.9558282208588957,
+      "grad_norm": 41.96194349470727,
+      "learning_rate": 5.507438016528925e-07,
+      "logits/chosen": -0.1640625,
+      "logits/rejected": -0.34375,
+      "logps/chosen": -434.0,
+      "logps/rejected": -348.0,
+      "loss": 0.2827,
+      "rewards/accuracies": 0.8203125,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -5.96875,
+      "step": 779
+    },
+    {
+      "epoch": 0.9570552147239264,
+      "grad_norm": 35.73940962793884,
+      "learning_rate": 5.504132231404958e-07,
+      "logits/chosen": -0.059326171875,
+      "logits/rejected": -0.2353515625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3415,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.2138671875,
+      "rewards/margins": 5.65625,
+      "rewards/rejected": -5.4375,
+      "step": 780
+    },
+    {
+      "epoch": 0.9582822085889571,
+      "grad_norm": 26.50283260922046,
+      "learning_rate": 5.500826446280991e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.408203125,
+      "logps/chosen": -418.0,
+      "logps/rejected": -342.0,
+      "loss": 0.1906,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 8.0625,
+      "rewards/rejected": -6.4375,
+      "step": 781
+    },
+    {
+      "epoch": 0.9595092024539877,
+      "grad_norm": 29.07388144480791,
+      "learning_rate": 5.497520661157025e-07,
+      "logits/chosen": -0.1376953125,
+      "logits/rejected": -0.2490234375,
+      "logps/chosen": -360.0,
+      "logps/rejected": -338.0,
+      "loss": 0.2204,
+      "rewards/accuracies": 0.890625,
+      "rewards/chosen": 0.875,
+      "rewards/margins": 6.4375,
+      "rewards/rejected": -5.5625,
+      "step": 782
+    },
+    {
+      "epoch": 0.9607361963190184,
+      "grad_norm": 36.281519570454215,
+      "learning_rate": 5.494214876033058e-07,
+      "logits/chosen": -0.1552734375,
+      "logits/rejected": -0.3046875,
+      "logps/chosen": -412.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2544,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.25,
+      "rewards/margins": 7.34375,
+      "rewards/rejected": -6.09375,
+      "step": 783
+    },
+    {
+      "epoch": 0.9619631901840491,
+      "grad_norm": 41.7797714543972,
+      "learning_rate": 5.490909090909091e-07,
+      "logits/chosen": -0.11181640625,
+      "logits/rejected": -0.162109375,
+      "logps/chosen": -378.0,
+      "logps/rejected": -340.0,
+      "loss": 0.3585,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 0.1767578125,
+      "rewards/margins": 5.8125,
+      "rewards/rejected": -5.625,
+      "step": 784
+    },
+    {
+      "epoch": 0.9631901840490797,
+      "grad_norm": 31.81554374340158,
+      "learning_rate": 5.487603305785124e-07,
+      "logits/chosen": -0.0751953125,
+      "logits/rejected": -0.24609375,
+      "logps/chosen": -374.0,
+      "logps/rejected": -308.0,
+      "loss": 0.2472,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.88671875,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -5.8125,
+      "step": 785
+    },
+    {
+      "epoch": 0.9644171779141104,
+      "grad_norm": 27.670018917456815,
+      "learning_rate": 5.484297520661156e-07,
+      "logits/chosen": -0.0732421875,
+      "logits/rejected": -0.28515625,
+      "logps/chosen": -364.0,
+      "logps/rejected": -318.0,
+      "loss": 0.2222,
+      "rewards/accuracies": 0.921875,
+      "rewards/chosen": 0.4921875,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -6.25,
+      "step": 786
+    },
+    {
+      "epoch": 0.9656441717791411,
+      "grad_norm": 37.16124889650476,
+      "learning_rate": 5.480991735537189e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.2890625,
+      "logps/chosen": -404.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2838,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.69140625,
+      "rewards/margins": 7.125,
+      "rewards/rejected": -6.4375,
+      "step": 787
+    },
+    {
+      "epoch": 0.9668711656441717,
+      "grad_norm": 24.391962735004444,
+      "learning_rate": 5.477685950413223e-07,
+      "logits/chosen": -0.1708984375,
+      "logits/rejected": -0.3515625,
+      "logps/chosen": -382.0,
+      "logps/rejected": -352.0,
+      "loss": 0.1706,
+      "rewards/accuracies": 0.9296875,
+      "rewards/chosen": 1.2734375,
+      "rewards/margins": 7.53125,
+      "rewards/rejected": -6.25,
+      "step": 788
+    },
+    {
+      "epoch": 0.9680981595092024,
+      "grad_norm": 32.024862187416,
+      "learning_rate": 5.474380165289256e-07,
+      "logits/chosen": -0.07958984375,
+      "logits/rejected": -0.265625,
+      "logps/chosen": -374.0,
+      "logps/rejected": -316.0,
+      "loss": 0.2843,
+      "rewards/accuracies": 0.8671875,
+      "rewards/chosen": 0.408203125,
+      "rewards/margins": 6.1875,
+      "rewards/rejected": -5.78125,
+      "step": 789
+    },
+    {
+      "epoch": 0.9693251533742331,
+      "grad_norm": 27.51754897072937,
+      "learning_rate": 5.471074380165289e-07,
+      "logits/chosen": -0.2041015625,
+      "logits/rejected": -0.423828125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -316.0,
+      "loss": 0.1866,
+      "rewards/accuracies": 0.90625,
+      "rewards/chosen": 0.70703125,
+      "rewards/margins": 6.71875,
+      "rewards/rejected": -6.03125,
+      "step": 790
+    },
+    {
+      "epoch": 0.9705521472392638,
+      "grad_norm": 37.440979455018265,
+      "learning_rate": 5.467768595041322e-07,
+      "logits/chosen": -0.1796875,
+      "logits/rejected": -0.392578125,
+      "logps/chosen": -390.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2983,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.1875,
+      "rewards/margins": 7.59375,
+      "rewards/rejected": -6.375,
+      "step": 791
+    },
+    {
+      "epoch": 0.9717791411042945,
+      "grad_norm": 36.704827910830204,
+      "learning_rate": 5.464462809917355e-07,
+      "logits/chosen": -0.1728515625,
+      "logits/rejected": -0.32421875,
+      "logps/chosen": -430.0,
+      "logps/rejected": -360.0,
+      "loss": 0.3431,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.2265625,
+      "rewards/margins": 6.25,
+      "rewards/rejected": -5.03125,
+      "step": 792
+    },
+    {
+      "epoch": 0.9730061349693252,
+      "grad_norm": 37.60757704339766,
+      "learning_rate": 5.461157024793389e-07,
+      "logits/chosen": -0.15234375,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3237,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 0.79296875,
+      "rewards/margins": 6.21875,
+      "rewards/rejected": -5.4375,
+      "step": 793
+    },
+    {
+      "epoch": 0.9742331288343559,
+      "grad_norm": 37.59792190871499,
+      "learning_rate": 5.457851239669422e-07,
+      "logits/chosen": -0.103515625,
+      "logits/rejected": -0.2431640625,
+      "logps/chosen": -350.0,
+      "logps/rejected": -322.0,
+      "loss": 0.3909,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.21875,
+      "rewards/margins": 6.0,
+      "rewards/rejected": -5.78125,
+      "step": 794
+    },
+    {
+      "epoch": 0.9754601226993865,
+      "grad_norm": 28.85691430451381,
+      "learning_rate": 5.454545454545454e-07,
+      "logits/chosen": -0.291015625,
+      "logits/rejected": -0.466796875,
+      "logps/chosen": -384.0,
+      "logps/rejected": -356.0,
+      "loss": 0.2394,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.59375,
+      "rewards/margins": 8.5,
+      "rewards/rejected": -6.90625,
+      "step": 795
+    },
+    {
+      "epoch": 0.9766871165644172,
+      "grad_norm": 41.11938718622486,
+      "learning_rate": 5.451239669421487e-07,
+      "logits/chosen": -0.189453125,
+      "logits/rejected": -0.400390625,
+      "logps/chosen": -402.0,
+      "logps/rejected": -322.0,
+      "loss": 0.2984,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": -0.0673828125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -6.375,
+      "step": 796
+    },
+    {
+      "epoch": 0.9779141104294479,
+      "grad_norm": 30.942308293626358,
+      "learning_rate": 5.44793388429752e-07,
+      "logits/chosen": -0.1171875,
+      "logits/rejected": -0.30078125,
+      "logps/chosen": -396.0,
+      "logps/rejected": -308.0,
+      "loss": 0.254,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.6953125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.625,
+      "step": 797
+    },
+    {
+      "epoch": 0.9791411042944785,
+      "grad_norm": 36.786308023977305,
+      "learning_rate": 5.444628099173553e-07,
+      "logits/chosen": -0.154296875,
+      "logits/rejected": -0.33984375,
+      "logps/chosen": -424.0,
+      "logps/rejected": -362.0,
+      "loss": 0.2823,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 0.76953125,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -6.125,
+      "step": 798
+    },
+    {
+      "epoch": 0.9803680981595092,
+      "grad_norm": 40.89607573095922,
+      "learning_rate": 5.441322314049587e-07,
+      "logits/chosen": -0.1953125,
+      "logits/rejected": -0.369140625,
+      "logps/chosen": -392.0,
+      "logps/rejected": -346.0,
+      "loss": 0.3222,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.109375,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -5.78125,
+      "step": 799
+    },
+    {
+      "epoch": 0.9815950920245399,
+      "grad_norm": 36.10519822526459,
+      "learning_rate": 5.43801652892562e-07,
+      "logits/chosen": -0.173828125,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -378.0,
+      "logps/rejected": -330.0,
+      "loss": 0.2658,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 0.61328125,
+      "rewards/margins": 6.625,
+      "rewards/rejected": -6.0,
+      "step": 800
+    },
+    {
+      "epoch": 0.9828220858895705,
+      "grad_norm": 42.48783397210697,
+      "learning_rate": 5.434710743801653e-07,
+      "logits/chosen": -0.17578125,
+      "logits/rejected": -0.3359375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -366.0,
+      "loss": 0.3944,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": 0.765625,
+      "rewards/margins": 5.9375,
+      "rewards/rejected": -5.1875,
+      "step": 801
+    },
+    {
+      "epoch": 0.9840490797546012,
+      "grad_norm": 43.552763355211724,
+      "learning_rate": 5.431404958677686e-07,
+      "logits/chosen": -0.2255859375,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3899,
+      "rewards/accuracies": 0.8046875,
+      "rewards/chosen": 0.453125,
+      "rewards/margins": 6.3125,
+      "rewards/rejected": -5.84375,
+      "step": 802
+    },
+    {
+      "epoch": 0.9852760736196319,
+      "grad_norm": 32.09849729457811,
+      "learning_rate": 5.428099173553719e-07,
+      "logits/chosen": -0.287109375,
+      "logits/rejected": -0.45703125,
+      "logps/chosen": -392.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2997,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.1328125,
+      "rewards/margins": 6.90625,
+      "rewards/rejected": -5.78125,
+      "step": 803
+    },
+    {
+      "epoch": 0.9865030674846625,
+      "grad_norm": 32.33615160143916,
+      "learning_rate": 5.424793388429751e-07,
+      "logits/chosen": -0.11962890625,
+      "logits/rejected": -0.240234375,
+      "logps/chosen": -352.0,
+      "logps/rejected": -328.0,
+      "loss": 0.3208,
+      "rewards/accuracies": 0.8515625,
+      "rewards/chosen": 0.251953125,
+      "rewards/margins": 5.90625,
+      "rewards/rejected": -5.65625,
+      "step": 804
+    },
+    {
+      "epoch": 0.9877300613496932,
+      "grad_norm": 39.61512778302868,
+      "learning_rate": 5.421487603305785e-07,
+      "logits/chosen": -0.224609375,
+      "logits/rejected": -0.37109375,
+      "logps/chosen": -428.0,
+      "logps/rejected": -380.0,
+      "loss": 0.3968,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": 1.0234375,
+      "rewards/margins": 6.96875,
+      "rewards/rejected": -5.96875,
+      "step": 805
+    },
+    {
+      "epoch": 0.9889570552147239,
+      "grad_norm": 40.74080836640833,
+      "learning_rate": 5.418181818181818e-07,
+      "logits/chosen": -0.01202392578125,
+      "logits/rejected": -0.1474609375,
+      "logps/chosen": -408.0,
+      "logps/rejected": -368.0,
+      "loss": 0.3811,
+      "rewards/accuracies": 0.7890625,
+      "rewards/chosen": 0.8046875,
+      "rewards/margins": 6.15625,
+      "rewards/rejected": -5.34375,
+      "step": 806
+    },
+    {
+      "epoch": 0.9901840490797545,
+      "grad_norm": 36.652039792723684,
+      "learning_rate": 5.414876033057851e-07,
+      "logits/chosen": -0.1005859375,
+      "logits/rejected": -0.26171875,
+      "logps/chosen": -394.0,
+      "logps/rejected": -334.0,
+      "loss": 0.2327,
+      "rewards/accuracies": 0.8828125,
+      "rewards/chosen": 1.078125,
+      "rewards/margins": 7.21875,
+      "rewards/rejected": -6.125,
+      "step": 807
+    },
+    {
+      "epoch": 0.9914110429447853,
+      "grad_norm": 33.57961890786831,
+      "learning_rate": 5.411570247933884e-07,
+      "logits/chosen": -0.2353515625,
+      "logits/rejected": -0.2578125,
+      "logps/chosen": -358.0,
+      "logps/rejected": -336.0,
+      "loss": 0.3091,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 0.11572265625,
+      "rewards/margins": 6.65625,
+      "rewards/rejected": -6.53125,
+      "step": 808
+    },
+    {
+      "epoch": 0.992638036809816,
+      "grad_norm": 35.60132202210017,
+      "learning_rate": 5.408264462809917e-07,
+      "logits/chosen": -0.1533203125,
+      "logits/rejected": -0.3828125,
+      "logps/chosen": -430.0,
+      "logps/rejected": -344.0,
+      "loss": 0.2618,
+      "rewards/accuracies": 0.828125,
+      "rewards/chosen": 1.609375,
+      "rewards/margins": 7.875,
+      "rewards/rejected": -6.28125,
+      "step": 809
+    },
+    {
+      "epoch": 0.9938650306748467,
+      "grad_norm": 33.977380429457895,
+      "learning_rate": 5.40495867768595e-07,
+      "logits/chosen": -0.251953125,
+      "logits/rejected": -0.4375,
+      "logps/chosen": -438.0,
+      "logps/rejected": -368.0,
+      "loss": 0.227,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.21875,
+      "rewards/margins": 7.28125,
+      "rewards/rejected": -6.0625,
+      "step": 810
+    },
+    {
+      "epoch": 0.9950920245398773,
+      "grad_norm": 31.635546850530186,
+      "learning_rate": 5.401652892561984e-07,
+      "logits/chosen": -0.234375,
+      "logits/rejected": -0.38671875,
+      "logps/chosen": -386.0,
+      "logps/rejected": -366.0,
+      "loss": 0.2741,
+      "rewards/accuracies": 0.859375,
+      "rewards/chosen": 1.0703125,
+      "rewards/margins": 6.78125,
+      "rewards/rejected": -5.71875,
+      "step": 811
+    },
+    {
+      "epoch": 0.996319018404908,
+      "grad_norm": 40.80563506539147,
+      "learning_rate": 5.398347107438017e-07,
+      "logits/chosen": -0.2119140625,
+      "logits/rejected": -0.294921875,
+      "logps/chosen": -358.0,
+      "logps/rejected": -328.0,
+      "loss": 0.2687,
+      "rewards/accuracies": 0.9140625,
+      "rewards/chosen": 0.4609375,
+      "rewards/margins": 7.0,
+      "rewards/rejected": -6.53125,
+      "step": 812
+    },
+    {
+      "epoch": 0.9975460122699387,
+      "grad_norm": 30.19996782105712,
+      "learning_rate": 5.395041322314049e-07,
+      "logits/chosen": -0.1337890625,
+      "logits/rejected": -0.279296875,
+      "logps/chosen": -338.0,
+      "logps/rejected": -324.0,
+      "loss": 0.1913,
+      "rewards/accuracies": 0.8984375,
+      "rewards/chosen": 0.87109375,
+      "rewards/margins": 7.0625,
+      "rewards/rejected": -6.1875,
+      "step": 813
+    },
+    {
+      "epoch": 0.9987730061349693,
+      "grad_norm": 37.73036986625575,
+      "learning_rate": 5.391735537190082e-07,
+      "logits/chosen": -0.11767578125,
+      "logits/rejected": -0.201171875,
+      "logps/chosen": -380.0,
+      "logps/rejected": -358.0,
+      "loss": 0.3425,
+      "rewards/accuracies": 0.8359375,
+      "rewards/chosen": -0.1416015625,
+      "rewards/margins": 5.5,
+      "rewards/rejected": -5.65625,
+      "step": 814
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 40.870567834511405,
+      "learning_rate": 5.388429752066115e-07,
+      "logits/chosen": -0.0458984375,
+      "logits/rejected": -0.23828125,
+      "logps/chosen": -380.0,
+      "logps/rejected": -350.0,
+      "loss": 0.2125,
+      "rewards/accuracies": 0.8571429252624512,
+      "rewards/chosen": 0.59375,
+      "rewards/margins": 6.875,
+      "rewards/rejected": -6.28125,
+      "step": 815
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2445,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-815/training_args.bin b/checkpoint-815/training_args.bin
new file mode 100644
index 0000000..70c7ec2
--- /dev/null
+++ b/checkpoint-815/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:528936595b6d96d00e537c7ae1257bb1454c1a31a8f47041035d37b1fc6d35e1
+size 9528
diff --git a/checkpoint-815/zero_to_fp32.py b/checkpoint-815/zero_to_fp32.py
new file mode 100644
index 0000000..24cc342
--- /dev/null
+++ b/checkpoint-815/zero_to_fp32.py
@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/config.json b/config.json
new file mode 100644
index 0000000..ded5a8b
--- /dev/null
+++ b/config.json
@@ -0,0 +1,35 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128009,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 8.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.50.0",
+  "use_cache": false,
+  "vocab_size": 128256
+}
diff --git a/configuration.json b/configuration.json
new file mode 100644
index 0000000..bbeeda1
--- /dev/null
+++ b/configuration.json
@@ -0,0 +1 @@
+{"framework": "pytorch", "task": "text-generation", "allow_remote": true}
\ No newline at end of file
diff --git a/generation_config.json b/generation_config.json
new file mode 100644
index 0000000..0acdb5a
--- /dev/null
+++ b/generation_config.json
@@ -0,0 +1,12 @@
+{
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.50.0"
+}
diff --git a/model-00001-of-00004.safetensors b/model-00001-of-00004.safetensors
new file mode 100644
index 0000000..8672400
--- /dev/null
+++ b/model-00001-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45f844c7bcec934a51bb941e54d30a655fc8db45dd739d36738f2fa4ce1de909
+size 4976698672
diff --git a/model-00002-of-00004.safetensors b/model-00002-of-00004.safetensors
new file mode 100644
index 0000000..5ef40e7
--- /dev/null
+++ b/model-00002-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e194d3a55358b5c5d86589c20ee9b494530bf727cc1955b7907610cd4b044682
+size 4999802720
diff --git a/model-00003-of-00004.safetensors b/model-00003-of-00004.safetensors
new file mode 100644
index 0000000..2715e04
--- /dev/null
+++ b/model-00003-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:774b89d412ab7fd49b257aa6ea2b15450e5a7caaff911791ce76707c059730b3
+size 4915916176
diff --git a/model-00004-of-00004.safetensors b/model-00004-of-00004.safetensors
new file mode 100644
index 0000000..f3b4e77
--- /dev/null
+++ b/model-00004-of-00004.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:933eca7410a1d89ab7601cb2b6a52d50488ac67c7c5bd30235b9802488e85d19
+size 1168138808
diff --git a/model.safetensors.index.json b/model.safetensors.index.json
new file mode 100644
index 0000000..0fd8120
--- /dev/null
+++ b/model.safetensors.index.json
@@ -0,0 +1,298 @@
+{
+  "metadata": {
+    "total_size": 16060522496
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00004-of-00004.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.norm.weight": "model-00004-of-00004.safetensors"
+  }
+}
diff --git a/runs/Apr03_06-04-06_ip-10-1-21-143/events.out.tfevents.1743660284.ip-10-1-21-143.961514.0 b/runs/Apr03_06-04-06_ip-10-1-21-143/events.out.tfevents.1743660284.ip-10-1-21-143.961514.0
new file mode 100644
index 0000000..1c23b69
--- /dev/null
+++ b/runs/Apr03_06-04-06_ip-10-1-21-143/events.out.tfevents.1743660284.ip-10-1-21-143.961514.0
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b12a72a16b77b95a2d9eddc26e8d638dddc291bdc42aa68844a884b61349f37c
+size 1688667
diff --git a/special_tokens_map.json b/special_tokens_map.json
new file mode 100644
index 0000000..278b7f0
--- /dev/null
+++ b/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/tokenizer.json b/tokenizer.json
new file mode 100644
index 0000000..1c1d8d5
--- /dev/null
+++ b/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920
diff --git a/tokenizer_config.json b/tokenizer_config.json
new file mode 100644
index 0000000..8098161
--- /dev/null
+++ b/tokenizer_config.json
@@ -0,0 +1,2064 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "<|end_of_text|>",
+  "tokenizer_class": "PreTrainedTokenizer"
+}
diff --git a/training_args.bin b/training_args.bin
new file mode 100644
index 0000000..70c7ec2
--- /dev/null
+++ b/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:528936595b6d96d00e537c7ae1257bb1454c1a31a8f47041035d37b1fc6d35e1
+size 9528