{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.9984, "eval_steps": 50, "global_step": 312, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0310546875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1433.4, "completions/mean_length": 119.246484375, "completions/mean_terminated_length": 73.85710296630859, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.016, "grad_norm": 0.0008150995709002018, "learning_rate": 3.1249999999999997e-07, "loss": -0.0017, "num_tokens": 12921436.0, "reward": 0.003515625, "reward_std": 0.017885054275393487, "rewards/accuracy_reward": 0.0013671875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.0056640625, "rewards/mean_confidence_reward": 0.0, "step": 5 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03466796875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1484.8, "completions/mean_length": 127.82548828125, "completions/mean_terminated_length": 77.27315979003906, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.032, "grad_norm": 0.0007543498068116605, "learning_rate": 6.249999999999999e-07, "loss": -0.0019, "num_tokens": 26187041.0, "reward": 0.00322265625, "reward_std": 0.016590457782149313, "rewards/accuracy_reward": 0.00048828125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.00595703125, "rewards/mean_confidence_reward": 0.0, "step": 10 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.040625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1381.6, "completions/mean_length": 139.8619140625, "completions/mean_terminated_length": 80.7786148071289, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.048, "grad_norm": 0.0032071403693407774, "learning_rate": 9.374999999999999e-07, "loss": -0.0029, "num_tokens": 39524283.0, "reward": 0.006591796875, "reward_std": 0.030955403670668603, "rewards/accuracy_reward": 0.00146484375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.01171875, "rewards/mean_confidence_reward": 0.0, "step": 15 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0451171875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1505.8, "completions/mean_length": 189.66513671875, "completions/mean_terminated_length": 126.14910278320312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.064, "grad_norm": 0.007881640456616879, "learning_rate": 1e-06, "loss": -0.001, "num_tokens": 53241174.0, "reward": 0.134912109375, "reward_std": 0.2078026980161667, "rewards/accuracy_reward": 0.039453125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.23037109375, "rewards/mean_confidence_reward": 0.0, "step": 20 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03349609375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1409.2, "completions/mean_length": 193.50859375, "completions/mean_terminated_length": 147.04785766601563, "completions/min_length": 7.4, "completions/min_terminated_length": 7.4, "epoch": 0.08, "grad_norm": 0.0024905798491090536, "learning_rate": 1e-06, "loss": 0.0459, "num_tokens": 67012174.0, "reward": 0.455810546875, "reward_std": 0.2630597770214081, "rewards/accuracy_reward": 0.13310546875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.778515625, "rewards/mean_confidence_reward": 0.0, "step": 25 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.02666015625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1414.0, "completions/mean_length": 166.709375, "completions/mean_terminated_length": 129.19416198730468, "completions/min_length": 12.8, "completions/min_terminated_length": 12.8, "epoch": 0.096, "grad_norm": 0.0015801023691892624, "learning_rate": 1e-06, "loss": 0.0386, "num_tokens": 80620206.0, "reward": 0.5583984375, "reward_std": 0.19487487077713012, "rewards/accuracy_reward": 0.1865234375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9302734375, "rewards/mean_confidence_reward": 0.0, "step": 30 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.01923828125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1417.8, "completions/mean_length": 142.8982421875, "completions/mean_terminated_length": 115.58807983398438, "completions/min_length": 16.8, "completions/min_terminated_length": 16.8, "epoch": 0.112, "grad_norm": 0.0015868425834923983, "learning_rate": 1e-06, "loss": 0.0273, "num_tokens": 94049276.0, "reward": 0.611572265625, "reward_std": 0.1569082885980606, "rewards/accuracy_reward": 0.2470703125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.97607421875, "rewards/mean_confidence_reward": 0.0, "step": 35 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0119140625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1156.2, "completions/mean_length": 122.52958984375, "completions/mean_terminated_length": 105.50102081298829, "completions/min_length": 18.4, "completions/min_terminated_length": 18.4, "epoch": 0.128, "grad_norm": 0.0013402927434071898, "learning_rate": 1e-06, "loss": 0.0172, "num_tokens": 107076971.0, "reward": 0.623876953125, "reward_std": 0.14180577099323272, "rewards/accuracy_reward": 0.26103515625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.98671875, "rewards/mean_confidence_reward": 0.0, "step": 40 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0072265625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1047.0, "completions/mean_length": 110.994921875, "completions/mean_terminated_length": 100.62855377197266, "completions/min_length": 21.8, "completions/min_terminated_length": 21.8, "epoch": 0.144, "grad_norm": 0.001353309373371303, "learning_rate": 1e-06, "loss": 0.0126, "num_tokens": 120020311.0, "reward": 0.663330078125, "reward_std": 0.14958325326442717, "rewards/accuracy_reward": 0.334375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99228515625, "rewards/mean_confidence_reward": 0.0, "step": 45 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00498046875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1120.8, "completions/mean_length": 111.87373046875, "completions/mean_terminated_length": 104.74790496826172, "completions/min_length": 24.8, "completions/min_terminated_length": 24.8, "epoch": 0.16, "grad_norm": 0.0012012316146865487, "learning_rate": 1e-06, "loss": 0.0074, "num_tokens": 133043146.0, "reward": 0.665576171875, "reward_std": 0.1308148130774498, "rewards/accuracy_reward": 0.336328125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99482421875, "rewards/mean_confidence_reward": 0.0, "step": 50 }, { "epoch": 0.16, "eval_completions/clipped_ratio": 0.00390625, "eval_completions/max_length": 938.0, "eval_completions/max_terminated_length": 375.0, "eval_completions/mean_length": 109.47684860229492, "eval_completions/mean_terminated_length": 103.89492416381836, "eval_completions/min_length": 36.5, "eval_completions/min_terminated_length": 36.5, "eval_loss": 0.0, "eval_num_tokens": 133043146.0, "eval_reward": 0.6416015625, "eval_reward_std": 0.2293500304222107, "eval_rewards/accuracy_reward": 0.287109375, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 0.99609375, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 20.8971, "eval_samples_per_second": 23.927, "eval_steps_per_second": 0.096, "step": 50 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00185546875, "completions/max_length": 1536.0, "completions/max_terminated_length": 812.2, "completions/mean_length": 106.87470703125, "completions/mean_terminated_length": 104.2185287475586, "completions/min_length": 27.4, "completions/min_terminated_length": 27.4, "epoch": 0.176, "grad_norm": 0.0012836462119594216, "learning_rate": 1e-06, "loss": 0.0042, "num_tokens": 146230983.0, "reward": 0.668994140625, "reward_std": 0.1268217846751213, "rewards/accuracy_reward": 0.33994140625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.998046875, "rewards/mean_confidence_reward": 0.0, "step": 55 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00126953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 887.6, "completions/mean_length": 108.08828125, "completions/mean_terminated_length": 106.27346343994141, "completions/min_length": 26.4, "completions/min_terminated_length": 26.4, "epoch": 0.192, "grad_norm": 0.001086747506633401, "learning_rate": 1e-06, "loss": 0.0029, "num_tokens": 159008943.0, "reward": 0.692236328125, "reward_std": 0.12218783050775528, "rewards/accuracy_reward": 0.38603515625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9984375, "rewards/mean_confidence_reward": 0.0, "step": 60 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00087890625, "completions/max_length": 1419.0, "completions/max_terminated_length": 922.8, "completions/mean_length": 115.68837890625, "completions/mean_terminated_length": 114.4382110595703, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "epoch": 0.208, "grad_norm": 0.0010557821951806545, "learning_rate": 1e-06, "loss": 0.0023, "num_tokens": 172082136.0, "reward": 0.702490234375, "reward_std": 0.12264649122953415, "rewards/accuracy_reward": 0.40615234375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.998828125, "rewards/mean_confidence_reward": 0.0, "step": 65 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00107421875, "completions/max_length": 1381.4, "completions/max_terminated_length": 742.2, "completions/mean_length": 120.01796875, "completions/mean_terminated_length": 118.49598541259766, "completions/min_length": 31.6, "completions/min_terminated_length": 31.6, "epoch": 0.224, "grad_norm": 0.001011895015835762, "learning_rate": 1e-06, "loss": 0.0026, "num_tokens": 185320624.0, "reward": 0.69443359375, "reward_std": 0.11083863973617554, "rewards/accuracy_reward": 0.39013671875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99873046875, "rewards/mean_confidence_reward": 0.0, "step": 70 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00068359375, "completions/max_length": 1536.0, "completions/max_terminated_length": 876.8, "completions/mean_length": 120.59326171875, "completions/mean_terminated_length": 119.62486572265625, "completions/min_length": 27.6, "completions/min_terminated_length": 27.6, "epoch": 0.24, "grad_norm": 0.0010586382122710347, "learning_rate": 1e-06, "loss": 0.0026, "num_tokens": 198663499.0, "reward": 0.71796875, "reward_std": 0.12902136147022247, "rewards/accuracy_reward": 0.437109375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.998828125, "rewards/mean_confidence_reward": 0.0, "step": 75 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00078125, "completions/max_length": 1335.8, "completions/max_terminated_length": 729.8, "completions/mean_length": 119.75810546875, "completions/mean_terminated_length": 118.650390625, "completions/min_length": 32.4, "completions/min_terminated_length": 32.4, "epoch": 0.256, "grad_norm": 0.0010895554441958666, "learning_rate": 1e-06, "loss": 0.0027, "num_tokens": 211800958.0, "reward": 0.701953125, "reward_std": 0.11856168359518052, "rewards/accuracy_reward": 0.40517578125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99873046875, "rewards/mean_confidence_reward": 0.0, "step": 80 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00078125, "completions/max_length": 1317.8, "completions/max_terminated_length": 551.0, "completions/mean_length": 121.72509765625, "completions/mean_terminated_length": 120.61879272460938, "completions/min_length": 32.6, "completions/min_terminated_length": 32.6, "epoch": 0.272, "grad_norm": 0.0009086201316677034, "learning_rate": 1e-06, "loss": 0.0022, "num_tokens": 224869439.0, "reward": 0.7064453125, "reward_std": 0.1061361700296402, "rewards/accuracy_reward": 0.41376953125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99912109375, "rewards/mean_confidence_reward": 0.0, "step": 85 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0009765625, "completions/max_length": 1536.0, "completions/max_terminated_length": 558.8, "completions/mean_length": 114.6669921875, "completions/mean_terminated_length": 113.27888793945313, "completions/min_length": 32.0, "completions/min_terminated_length": 32.0, "epoch": 0.288, "grad_norm": 0.0010958426864817739, "learning_rate": 1e-06, "loss": 0.0025, "num_tokens": 237858125.0, "reward": 0.71142578125, "reward_std": 0.11188835352659225, "rewards/accuracy_reward": 0.42412109375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99873046875, "rewards/mean_confidence_reward": 0.0, "step": 90 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1094.8, "completions/max_terminated_length": 696.6, "completions/mean_length": 112.71875, "completions/mean_terminated_length": 112.02296142578125, "completions/min_length": 33.0, "completions/min_terminated_length": 33.0, "epoch": 0.304, "grad_norm": 0.0011008486617356539, "learning_rate": 1e-06, "loss": 0.0016, "num_tokens": 250798637.0, "reward": 0.72265625, "reward_std": 0.10946107506752015, "rewards/accuracy_reward": 0.4458984375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 95 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0009765625, "completions/max_length": 1536.0, "completions/max_terminated_length": 751.6, "completions/mean_length": 114.919921875, "completions/mean_terminated_length": 113.53117523193359, "completions/min_length": 33.2, "completions/min_terminated_length": 33.2, "epoch": 0.32, "grad_norm": 0.0009863615268841386, "learning_rate": 1e-06, "loss": 0.0024, "num_tokens": 263920441.0, "reward": 0.725830078125, "reward_std": 0.09988638162612914, "rewards/accuracy_reward": 0.452734375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99892578125, "rewards/mean_confidence_reward": 0.0, "step": 100 }, { "epoch": 0.32, "eval_completions/clipped_ratio": 0.0, "eval_completions/max_length": 466.0, "eval_completions/max_terminated_length": 466.0, "eval_completions/mean_length": 116.24256896972656, "eval_completions/mean_terminated_length": 116.24256896972656, "eval_completions/min_length": 41.0, "eval_completions/min_terminated_length": 41.0, "eval_loss": 0.0, "eval_num_tokens": 263920441.0, "eval_reward": 0.6689453125, "eval_reward_std": 0.23459484428167343, "eval_rewards/accuracy_reward": 0.337890625, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 1.0, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 12.6225, "eval_samples_per_second": 39.612, "eval_steps_per_second": 0.158, "step": 100 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1208.6, "completions/max_terminated_length": 697.0, "completions/mean_length": 115.3330078125, "completions/mean_terminated_length": 114.63796844482422, "completions/min_length": 29.2, "completions/min_terminated_length": 29.2, "epoch": 0.336, "grad_norm": 0.0010649055475369096, "learning_rate": 1e-06, "loss": 0.0018, "num_tokens": 276680203.0, "reward": 0.728515625, "reward_std": 0.10526287704706191, "rewards/accuracy_reward": 0.4580078125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9990234375, "rewards/mean_confidence_reward": 0.0, "step": 105 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00078125, "completions/max_length": 1176.8, "completions/max_terminated_length": 753.2, "completions/mean_length": 117.95869140625, "completions/mean_terminated_length": 116.84952392578126, "completions/min_length": 33.6, "completions/min_terminated_length": 33.6, "epoch": 0.352, "grad_norm": 0.000986444647423923, "learning_rate": 1e-06, "loss": 0.0023, "num_tokens": 290004836.0, "reward": 0.707763671875, "reward_std": 0.0988370344042778, "rewards/accuracy_reward": 0.41650390625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9990234375, "rewards/mean_confidence_reward": 0.0, "step": 110 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0009765625, "completions/max_length": 1105.8, "completions/max_terminated_length": 712.6, "completions/mean_length": 117.285546875, "completions/mean_terminated_length": 115.89883575439453, "completions/min_length": 36.6, "completions/min_terminated_length": 36.6, "epoch": 0.368, "grad_norm": 0.0009329163003712893, "learning_rate": 1e-06, "loss": 0.0019, "num_tokens": 303127632.0, "reward": 0.714794921875, "reward_std": 0.09199287295341492, "rewards/accuracy_reward": 0.43076171875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.998828125, "rewards/mean_confidence_reward": 0.0, "step": 115 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 737.8, "completions/max_terminated_length": 541.2, "completions/mean_length": 114.44951171875, "completions/mean_terminated_length": 113.8926025390625, "completions/min_length": 37.0, "completions/min_terminated_length": 37.0, "epoch": 0.384, "grad_norm": 0.001053647487424314, "learning_rate": 1e-06, "loss": 0.0009, "num_tokens": 316012427.0, "reward": 0.727099609375, "reward_std": 0.0868588350713253, "rewards/accuracy_reward": 0.4546875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99951171875, "rewards/mean_confidence_reward": 0.0, "step": 120 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0005859375, "completions/max_length": 1166.4, "completions/max_terminated_length": 508.4, "completions/mean_length": 116.58603515625, "completions/mean_terminated_length": 115.75442199707031, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "epoch": 0.4, "grad_norm": 0.0011356452014297247, "learning_rate": 1e-06, "loss": 0.0014, "num_tokens": 329099036.0, "reward": 0.719482421875, "reward_std": 0.09846941083669662, "rewards/accuracy_reward": 0.43974609375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99921875, "rewards/mean_confidence_reward": 0.0, "step": 125 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1332.6, "completions/max_terminated_length": 760.6, "completions/mean_length": 119.96689453125, "completions/mean_terminated_length": 119.41312713623047, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "epoch": 0.416, "grad_norm": 0.0009609300759620965, "learning_rate": 1e-06, "loss": 0.0013, "num_tokens": 342065001.0, "reward": 0.72802734375, "reward_std": 0.09362150728702545, "rewards/accuracy_reward": 0.45673828125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99931640625, "rewards/mean_confidence_reward": 0.0, "step": 130 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 9.765625e-05, "completions/max_length": 772.8, "completions/max_terminated_length": 560.2, "completions/mean_length": 122.2830078125, "completions/mean_terminated_length": 122.14535827636719, "completions/min_length": 43.4, "completions/min_terminated_length": 43.4, "epoch": 0.432, "grad_norm": 0.0009893623646348715, "learning_rate": 1e-06, "loss": 0.0008, "num_tokens": 355187835.0, "reward": 0.73603515625, "reward_std": 0.08680803030729294, "rewards/accuracy_reward": 0.47216796875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99990234375, "rewards/mean_confidence_reward": 0.0, "step": 135 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 902.8, "completions/max_terminated_length": 634.4, "completions/mean_length": 122.67548828125, "completions/mean_terminated_length": 122.3994155883789, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "epoch": 0.448, "grad_norm": 0.0009984844364225864, "learning_rate": 1e-06, "loss": 0.0011, "num_tokens": 368253152.0, "reward": 0.725439453125, "reward_std": 0.0879106342792511, "rewards/accuracy_reward": 0.451171875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 140 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1109.0, "completions/max_terminated_length": 534.2, "completions/mean_length": 129.035546875, "completions/mean_terminated_length": 128.35015716552735, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "epoch": 0.464, "grad_norm": 0.0008535922970622778, "learning_rate": 1e-06, "loss": 0.0012, "num_tokens": 381601612.0, "reward": 0.69501953125, "reward_std": 0.08753804713487626, "rewards/accuracy_reward": 0.39052734375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99951171875, "rewards/mean_confidence_reward": 0.0, "step": 145 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00107421875, "completions/max_length": 1536.0, "completions/max_terminated_length": 721.4, "completions/mean_length": 131.50673828125, "completions/mean_terminated_length": 129.99645080566407, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "epoch": 0.48, "grad_norm": 0.0010002210037782788, "learning_rate": 1e-06, "loss": 0.0026, "num_tokens": 394852593.0, "reward": 0.73046875, "reward_std": 0.09515460133552552, "rewards/accuracy_reward": 0.462109375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.998828125, "rewards/mean_confidence_reward": 0.0, "step": 150 }, { "epoch": 0.48, "eval_completions/clipped_ratio": 0.0, "eval_completions/max_length": 445.0, "eval_completions/max_terminated_length": 445.0, "eval_completions/mean_length": 135.29373168945312, "eval_completions/mean_terminated_length": 135.29373168945312, "eval_completions/min_length": 46.5, "eval_completions/min_terminated_length": 46.5, "eval_loss": 0.0, "eval_num_tokens": 394852593.0, "eval_reward": 0.6796875, "eval_reward_std": 0.2372811809182167, "eval_rewards/accuracy_reward": 0.359375, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 1.0, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 12.2325, "eval_samples_per_second": 40.875, "eval_steps_per_second": 0.163, "step": 150 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0009765625, "completions/max_length": 1536.0, "completions/max_terminated_length": 557.2, "completions/mean_length": 130.08017578125, "completions/mean_terminated_length": 128.7072326660156, "completions/min_length": 48.0, "completions/min_terminated_length": 48.0, "epoch": 0.496, "grad_norm": 0.0009576234733685851, "learning_rate": 1e-06, "loss": 0.0025, "num_tokens": 408348774.0, "reward": 0.735693359375, "reward_std": 0.09800407886505128, "rewards/accuracy_reward": 0.4724609375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99892578125, "rewards/mean_confidence_reward": 0.0, "step": 155 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 847.4, "completions/max_terminated_length": 447.8, "completions/mean_length": 122.45078125, "completions/mean_terminated_length": 122.17521820068359, "completions/min_length": 45.4, "completions/min_terminated_length": 45.4, "epoch": 0.512, "grad_norm": 0.0011876981006935239, "learning_rate": 1e-06, "loss": 0.0014, "num_tokens": 421604654.0, "reward": 0.738671875, "reward_std": 0.09699147194623947, "rewards/accuracy_reward": 0.477734375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 160 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00087890625, "completions/max_length": 1444.8, "completions/max_terminated_length": 686.4, "completions/mean_length": 121.19296875, "completions/mean_terminated_length": 119.94681091308594, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "epoch": 0.528, "grad_norm": 0.0011259722523391247, "learning_rate": 1e-06, "loss": 0.0024, "num_tokens": 434731526.0, "reward": 0.734033203125, "reward_std": 0.09081923365592956, "rewards/accuracy_reward": 0.4689453125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99912109375, "rewards/mean_confidence_reward": 0.0, "step": 165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00068359375, "completions/max_length": 1314.2, "completions/max_terminated_length": 527.6, "completions/mean_length": 117.2203125, "completions/mean_terminated_length": 116.24924011230469, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "epoch": 0.544, "grad_norm": 0.0012388949980959296, "learning_rate": 1e-06, "loss": 0.0018, "num_tokens": 447951766.0, "reward": 0.75224609375, "reward_std": 0.1028323695063591, "rewards/accuracy_reward": 0.5052734375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99921875, "rewards/mean_confidence_reward": 0.0, "step": 170 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1536.0, "completions/max_terminated_length": 628.6, "completions/mean_length": 114.80283203125, "completions/mean_terminated_length": 114.10854949951172, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "epoch": 0.56, "grad_norm": 0.001046418328769505, "learning_rate": 1e-06, "loss": 0.0014, "num_tokens": 460805075.0, "reward": 0.72294921875, "reward_std": 0.0901212826371193, "rewards/accuracy_reward": 0.44658203125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99931640625, "rewards/mean_confidence_reward": 0.0, "step": 175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 869.6, "completions/max_terminated_length": 495.6, "completions/mean_length": 112.623828125, "completions/mean_terminated_length": 111.92962493896485, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "epoch": 0.576, "grad_norm": 0.0010307738557457924, "learning_rate": 1e-06, "loss": 0.0009, "num_tokens": 474001287.0, "reward": 0.723779296875, "reward_std": 0.07322348654270172, "rewards/accuracy_reward": 0.4482421875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99931640625, "rewards/mean_confidence_reward": 0.0, "step": 180 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 902.2, "completions/max_terminated_length": 642.8, "completions/mean_length": 107.84658203125, "completions/mean_terminated_length": 107.567431640625, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "epoch": 0.592, "grad_norm": 0.0010062718065455556, "learning_rate": 1e-06, "loss": 0.0006, "num_tokens": 487129668.0, "reward": 0.727978515625, "reward_std": 0.07468962967395783, "rewards/accuracy_reward": 0.45634765625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 185 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 9.765625e-05, "completions/max_length": 644.2, "completions/max_terminated_length": 434.6, "completions/mean_length": 105.86552734375, "completions/mean_terminated_length": 105.72665405273438, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "epoch": 0.608, "grad_norm": 0.001235591946169734, "learning_rate": 1e-06, "loss": 0.0017, "num_tokens": 500069539.0, "reward": 0.7326171875, "reward_std": 0.08147054463624955, "rewards/accuracy_reward": 0.46533203125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99990234375, "rewards/mean_confidence_reward": 0.0, "step": 190 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1311.0, "completions/max_terminated_length": 410.4, "completions/mean_length": 100.38515625, "completions/mean_terminated_length": 99.68307495117188, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "epoch": 0.624, "grad_norm": 0.0012076563434675336, "learning_rate": 1e-06, "loss": 0.0014, "num_tokens": 513297707.0, "reward": 0.722314453125, "reward_std": 0.08004606068134308, "rewards/accuracy_reward": 0.4451171875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99951171875, "rewards/mean_confidence_reward": 0.0, "step": 195 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00078125, "completions/max_length": 1297.6, "completions/max_terminated_length": 421.6, "completions/mean_length": 100.12705078125, "completions/mean_terminated_length": 99.00437927246094, "completions/min_length": 35.2, "completions/min_terminated_length": 35.2, "epoch": 0.64, "grad_norm": 0.0009635328315198421, "learning_rate": 1e-06, "loss": 0.0014, "num_tokens": 526522016.0, "reward": 0.76123046875, "reward_std": 0.06789408177137375, "rewards/accuracy_reward": 0.5234375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9990234375, "rewards/mean_confidence_reward": 0.0, "step": 200 }, { "epoch": 0.64, "eval_completions/clipped_ratio": 0.0, "eval_completions/max_length": 349.0, "eval_completions/max_terminated_length": 349.0, "eval_completions/mean_length": 102.08308410644531, "eval_completions/mean_terminated_length": 102.08308410644531, "eval_completions/min_length": 43.0, "eval_completions/min_terminated_length": 43.0, "eval_loss": 0.0, "eval_num_tokens": 526522016.0, "eval_reward": 0.6982421875, "eval_reward_std": 0.24458713829517365, "eval_rewards/accuracy_reward": 0.396484375, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 1.0, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 10.5504, "eval_samples_per_second": 47.392, "eval_steps_per_second": 0.19, "step": 200 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 796.8, "completions/max_terminated_length": 636.6, "completions/mean_length": 101.0236328125, "completions/mean_terminated_length": 100.74494934082031, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "epoch": 0.656, "grad_norm": 0.0011409136932343245, "learning_rate": 1e-06, "loss": 0.0005, "num_tokens": 539269362.0, "reward": 0.7212890625, "reward_std": 0.0748053178191185, "rewards/accuracy_reward": 0.44287109375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 205 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1111.8, "completions/max_terminated_length": 482.0, "completions/mean_length": 102.7431640625, "completions/mean_terminated_length": 102.18379516601563, "completions/min_length": 36.6, "completions/min_terminated_length": 36.6, "epoch": 0.672, "grad_norm": 0.0010987328132614493, "learning_rate": 1e-06, "loss": 0.0006, "num_tokens": 552091212.0, "reward": 0.73291015625, "reward_std": 0.0707667201757431, "rewards/accuracy_reward": 0.4662109375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 210 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1298.6, "completions/max_terminated_length": 443.8, "completions/mean_length": 104.8025390625, "completions/mean_terminated_length": 104.10315551757813, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "epoch": 0.688, "grad_norm": 0.001303318771533668, "learning_rate": 1e-06, "loss": 0.0014, "num_tokens": 564974630.0, "reward": 0.737451171875, "reward_std": 0.08159364163875579, "rewards/accuracy_reward": 0.475390625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99951171875, "rewards/mean_confidence_reward": 0.0, "step": 215 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 9.765625e-05, "completions/max_length": 646.8, "completions/max_terminated_length": 413.4, "completions/mean_length": 103.31484375, "completions/mean_terminated_length": 103.17418212890625, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "epoch": 0.704, "grad_norm": 0.0012299970258027315, "learning_rate": 1e-06, "loss": 0.0004, "num_tokens": 577755038.0, "reward": 0.749462890625, "reward_std": 0.07313954383134842, "rewards/accuracy_reward": 0.49912109375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9998046875, "rewards/mean_confidence_reward": 0.0, "step": 220 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 9.765625e-05, "completions/max_length": 833.0, "completions/max_terminated_length": 596.4, "completions/mean_length": 99.44697265625, "completions/mean_terminated_length": 99.30665740966796, "completions/min_length": 30.2, "completions/min_terminated_length": 30.2, "epoch": 0.72, "grad_norm": 0.0013605753192678094, "learning_rate": 1e-06, "loss": 0.0007, "num_tokens": 590639551.0, "reward": 0.752490234375, "reward_std": 0.07478533685207367, "rewards/accuracy_reward": 0.505078125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99990234375, "rewards/mean_confidence_reward": 0.0, "step": 225 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0005859375, "completions/max_length": 1318.8, "completions/max_terminated_length": 461.4, "completions/mean_length": 101.459765625, "completions/mean_terminated_length": 100.61889038085937, "completions/min_length": 34.0, "completions/min_terminated_length": 34.0, "epoch": 0.736, "grad_norm": 0.001188904047012329, "learning_rate": 1e-06, "loss": 0.0009, "num_tokens": 603474403.0, "reward": 0.75400390625, "reward_std": 0.07088786214590073, "rewards/accuracy_reward": 0.50859375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 230 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0005859375, "completions/max_length": 1308.2, "completions/max_terminated_length": 705.2, "completions/mean_length": 108.83603515625, "completions/mean_terminated_length": 107.99874267578124, "completions/min_length": 35.4, "completions/min_terminated_length": 35.4, "epoch": 0.752, "grad_norm": 0.0011398772476240993, "learning_rate": 1e-06, "loss": 0.0008, "num_tokens": 616672404.0, "reward": 0.745458984375, "reward_std": 0.07853687852621079, "rewards/accuracy_reward": 0.4916015625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99931640625, "rewards/mean_confidence_reward": 0.0, "step": 235 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1314.2, "completions/max_terminated_length": 580.0, "completions/mean_length": 118.07548828125, "completions/mean_terminated_length": 117.52151641845703, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "epoch": 0.768, "grad_norm": 0.0011985410237684846, "learning_rate": 1e-06, "loss": 0.0007, "num_tokens": 629670521.0, "reward": 0.743798828125, "reward_std": 0.06558914035558701, "rewards/accuracy_reward": 0.4880859375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99951171875, "rewards/mean_confidence_reward": 0.0, "step": 240 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00078125, "completions/max_length": 1334.2, "completions/max_terminated_length": 662.0, "completions/mean_length": 123.54267578125, "completions/mean_terminated_length": 122.43762512207032, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "epoch": 0.784, "grad_norm": 0.0009077648282982409, "learning_rate": 1e-06, "loss": 0.0024, "num_tokens": 642966286.0, "reward": 0.753173828125, "reward_std": 0.08060203045606613, "rewards/accuracy_reward": 0.50732421875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9990234375, "rewards/mean_confidence_reward": 0.0, "step": 245 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00087890625, "completions/max_length": 1536.0, "completions/max_terminated_length": 708.6, "completions/mean_length": 129.51396484375, "completions/mean_terminated_length": 128.2775894165039, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "epoch": 0.8, "grad_norm": 0.0011181648587808013, "learning_rate": 1e-06, "loss": 0.0015, "num_tokens": 656159389.0, "reward": 0.756103515625, "reward_std": 0.0859486699104309, "rewards/accuracy_reward": 0.5130859375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99912109375, "rewards/mean_confidence_reward": 0.0, "step": 250 }, { "epoch": 0.8, "eval_completions/clipped_ratio": 0.0, "eval_completions/max_length": 378.5, "eval_completions/max_terminated_length": 378.5, "eval_completions/mean_length": 136.143310546875, "eval_completions/mean_terminated_length": 136.143310546875, "eval_completions/min_length": 55.5, "eval_completions/min_terminated_length": 55.5, "eval_loss": 0.0, "eval_num_tokens": 656159389.0, "eval_reward": 0.6982421875, "eval_reward_std": 0.2450297772884369, "eval_rewards/accuracy_reward": 0.396484375, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 1.0, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 11.1058, "eval_samples_per_second": 45.022, "eval_steps_per_second": 0.18, "step": 250 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0013671875, "completions/max_length": 1536.0, "completions/max_terminated_length": 624.4, "completions/mean_length": 126.7615234375, "completions/mean_terminated_length": 124.83155364990235, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "epoch": 0.816, "grad_norm": 0.0010861102491617203, "learning_rate": 1e-06, "loss": 0.0026, "num_tokens": 669412915.0, "reward": 0.762060546875, "reward_std": 0.08036677092313767, "rewards/accuracy_reward": 0.5255859375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99853515625, "rewards/mean_confidence_reward": 0.0, "step": 255 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 1049.4, "completions/max_terminated_length": 653.2, "completions/mean_length": 133.90185546875, "completions/mean_terminated_length": 133.627978515625, "completions/min_length": 45.6, "completions/min_terminated_length": 45.6, "epoch": 0.832, "grad_norm": 0.0010349162621423602, "learning_rate": 1e-06, "loss": 0.0005, "num_tokens": 682648742.0, "reward": 0.754443359375, "reward_std": 0.07279899269342423, "rewards/accuracy_reward": 0.5091796875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 260 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1366.8, "completions/max_terminated_length": 684.6, "completions/mean_length": 131.746484375, "completions/mean_terminated_length": 131.19830627441405, "completions/min_length": 48.8, "completions/min_terminated_length": 48.8, "epoch": 0.848, "grad_norm": 0.001083633047528565, "learning_rate": 1e-06, "loss": 0.0015, "num_tokens": 695868514.0, "reward": 0.750244140625, "reward_std": 0.07709781229496002, "rewards/accuracy_reward": 0.5009765625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99951171875, "rewards/mean_confidence_reward": 0.0, "step": 265 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00126953125, "completions/max_length": 1335.6, "completions/max_terminated_length": 843.4, "completions/mean_length": 139.9458984375, "completions/mean_terminated_length": 138.16998596191405, "completions/min_length": 54.0, "completions/min_terminated_length": 54.0, "epoch": 0.864, "grad_norm": 0.0009675953770056367, "learning_rate": 1e-06, "loss": 0.002, "num_tokens": 709144696.0, "reward": 0.760107421875, "reward_std": 0.0785179540514946, "rewards/accuracy_reward": 0.52158203125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9986328125, "rewards/mean_confidence_reward": 0.0, "step": 270 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0013671875, "completions/max_length": 1359.6, "completions/max_terminated_length": 809.4, "completions/mean_length": 140.3294921875, "completions/mean_terminated_length": 138.41428833007814, "completions/min_length": 52.0, "completions/min_terminated_length": 52.0, "epoch": 0.88, "grad_norm": 0.0010044885566458106, "learning_rate": 1e-06, "loss": 0.0021, "num_tokens": 722585062.0, "reward": 0.738671875, "reward_std": 0.06688929572701455, "rewards/accuracy_reward": 0.4787109375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9986328125, "rewards/mean_confidence_reward": 0.0, "step": 275 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00068359375, "completions/max_length": 1173.6, "completions/max_terminated_length": 641.0, "completions/mean_length": 135.87109375, "completions/mean_terminated_length": 134.9119842529297, "completions/min_length": 49.2, "completions/min_terminated_length": 49.2, "epoch": 0.896, "grad_norm": 0.0010312796803191304, "learning_rate": 1e-06, "loss": 0.0014, "num_tokens": 735943550.0, "reward": 0.748828125, "reward_std": 0.07459539771080018, "rewards/accuracy_reward": 0.49833984375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99931640625, "rewards/mean_confidence_reward": 0.0, "step": 280 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0005859375, "completions/max_length": 1287.0, "completions/max_terminated_length": 645.2, "completions/mean_length": 134.03544921875, "completions/mean_terminated_length": 133.21565856933594, "completions/min_length": 46.6, "completions/min_terminated_length": 46.6, "epoch": 0.912, "grad_norm": 0.0007955370820127428, "learning_rate": 1e-06, "loss": 0.0013, "num_tokens": 749223689.0, "reward": 0.740380859375, "reward_std": 0.06752934828400611, "rewards/accuracy_reward": 0.481640625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99912109375, "rewards/mean_confidence_reward": 0.0, "step": 285 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0009765625, "completions/max_length": 1536.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 130.778125, "completions/mean_terminated_length": 129.40672912597657, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "epoch": 0.928, "grad_norm": 0.000867258058860898, "learning_rate": 1e-06, "loss": 0.0012, "num_tokens": 762445993.0, "reward": 0.75517578125, "reward_std": 0.06989040896296501, "rewards/accuracy_reward": 0.511328125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9990234375, "rewards/mean_confidence_reward": 0.0, "step": 290 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0009765625, "completions/max_length": 1141.2, "completions/max_terminated_length": 584.0, "completions/mean_length": 128.4544921875, "completions/mean_terminated_length": 127.07853851318359, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "epoch": 0.944, "grad_norm": 0.0010988096473738551, "learning_rate": 1e-06, "loss": 0.0015, "num_tokens": 775593111.0, "reward": 0.752294921875, "reward_std": 0.07567534744739532, "rewards/accuracy_reward": 0.50556640625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9990234375, "rewards/mean_confidence_reward": 0.0, "step": 295 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1115.4, "completions/max_terminated_length": 573.2, "completions/mean_length": 130.62744140625, "completions/mean_terminated_length": 130.07823181152344, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "epoch": 0.96, "grad_norm": 0.0010277414694428444, "learning_rate": 1e-06, "loss": 0.0006, "num_tokens": 788727376.0, "reward": 0.740673828125, "reward_std": 0.06425100266933441, "rewards/accuracy_reward": 0.48212890625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99921875, "rewards/mean_confidence_reward": 0.0, "step": 300 }, { "epoch": 0.96, "eval_completions/clipped_ratio": 0.0, "eval_completions/max_length": 383.0, "eval_completions/max_terminated_length": 383.0, "eval_completions/mean_length": 138.22162628173828, "eval_completions/mean_terminated_length": 138.22162628173828, "eval_completions/min_length": 54.0, "eval_completions/min_terminated_length": 54.0, "eval_loss": 0.0, "eval_num_tokens": 788727376.0, "eval_reward": 0.708984375, "eval_reward_std": 0.2455444484949112, "eval_rewards/accuracy_reward": 0.41796875, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 1.0, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 11.3207, "eval_samples_per_second": 44.167, "eval_steps_per_second": 0.177, "step": 300 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0009765625, "completions/max_length": 1536.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 137.1462890625, "completions/mean_terminated_length": 135.7811706542969, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "epoch": 0.976, "grad_norm": 0.0010885481024160981, "learning_rate": 1e-06, "loss": 0.0014, "num_tokens": 801849194.0, "reward": 0.76015625, "reward_std": 0.07030820697546006, "rewards/accuracy_reward": 0.5212890625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9990234375, "rewards/mean_confidence_reward": 0.0, "step": 305 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1006.6, "completions/max_terminated_length": 640.2, "completions/mean_length": 131.81298828125, "completions/mean_terminated_length": 131.26700286865236, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "epoch": 0.992, "grad_norm": 0.0010889332043007016, "learning_rate": 1e-06, "loss": 0.0011, "num_tokens": 815183759.0, "reward": 0.750927734375, "reward_std": 0.06979047656059265, "rewards/accuracy_reward": 0.50244140625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 310 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 632.0, "completions/max_terminated_length": 632.0, "completions/mean_length": 127.36857604980469, "completions/mean_terminated_length": 127.36857604980469, "completions/min_length": 43.5, "completions/min_terminated_length": 43.5, "epoch": 0.9984, "num_tokens": 820458558.0, "reward": 0.7520751953125, "reward_std": 0.06880267150700092, "rewards/accuracy_reward": 0.50439453125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999755859375, "rewards/mean_confidence_reward": 0.0, "step": 312, "total_flos": 0.0, "train_loss": 0.003609725790458749, "train_runtime": 39158.0537, "train_samples_per_second": 0.511, "train_steps_per_second": 0.008 } ], "logging_steps": 5, "max_steps": 312, "num_input_tokens_seen": 820458558, "num_train_epochs": 1, "save_steps": 60, "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 }