commit 528f3c0db1ce5716c4158f58c8274300207e07a5
Author: ModelHub XC <noreply@modelhub.org.cn>
Date:   Sat Jun 13 09:56:16 2026 +0800

    初始化项目，由ModelHub XC社区提供模型
    
    Model: cs-552-2026-catma/multilingual_model
    Source: Original Platform

diff --git a/.gitattributes b/.gitattributes
new file mode 100644
index 0000000..670f935
--- /dev/null
+++ b/.gitattributes
@@ -0,0 +1,38 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt 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
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz 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
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl 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
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* 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
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+checkpoint-9000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-9329/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text
diff --git a/EVAL_REPORT.md b/EVAL_REPORT.md
new file mode 100644
index 0000000..ac485e7
--- /dev/null
+++ b/EVAL_REPORT.md
@@ -0,0 +1,53 @@
+# Automated MNLP evaluation report
+
+- **Model repo:** [`cs-552-2026-catma/multilingual_model`](https://huggingface.co/cs-552-2026-catma/multilingual_model)
+- **Owner(s):** group **catma**
+- **Generated at:** 2026-05-28T08:03:33+00:00 (UTC)
+- **Pipeline:** [mnlp-project-ci](https://github.com/eric11eca/mnlp-project-ci)
+
+_This PR is opened automatically by the course CI. It is **non-blocking** — you do not need to merge it. The next nightly run will refresh this file._
+
+## Summary
+
+| Benchmark | Accuracy | Status |
+|---|---:|---|
+| Math | — | not run |
+| Knowledge | — | not run |
+| Multilingual | 0.6500 | ok |
+| Safety | — | not run |
+
+## Sample completions
+
+_Prompts are intentionally omitted to avoid revealing benchmark contents. For multi-completion problems, only one completion is shown per sample._
+
+### Multilingual
+
+**Correct** (1 shown)
+
+- **reference**: `A`
+- **overall** (1/1 completions correct)
+- **extracted** (✓): `A`
+- **completion**:
+
+    ```text
+    <think>
+    Analizando las opciones, la respuesta correcta es A.
+    </think>
+    
+    \boxed{A}
+    ```
+
+**Incorrect** (1 shown)
+
+- **reference**: `B`
+- **overall** (0/1 completions correct)
+- **extracted** (✗): `A`
+- **completion**:
+
+    ```text
+    <think>
+    विकल्पों का विश्लेषण करते हुए, सही उत्तर A है।
+    </think>
+    
+    \boxed{A}
+    ```
diff --git a/README.md b/README.md
new file mode 100644
index 0000000..0ec7571
--- /dev/null
+++ b/README.md
@@ -0,0 +1,58 @@
+---
+base_model: Qwen/Qwen3-1.7B
+library_name: transformers
+model_name: multilingual
+tags:
+- generated_from_trainer
+- trl
+- sft
+licence: license
+---
+
+# Model Card for multilingual
+
+This model is a fine-tuned version of [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B).
+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
+
+ 
+
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 1.3.0
+- Transformers: 5.7.0
+- Pytorch: 2.10.0+cu128
+- Datasets: 4.8.5
+- Tokenizers: 0.22.2
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@software{vonwerra2020trl,
+  title   = {{TRL: Transformers Reinforcement Learning}},
+  author  = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
+  license = {Apache-2.0},
+  url     = {https://github.com/huggingface/trl},
+  year    = {2020}
+}
+```
\ No newline at end of file
diff --git a/chat_template.jinja b/chat_template.jinja
new file mode 100644
index 0000000..01be9b3
--- /dev/null
+++ b/chat_template.jinja
@@ -0,0 +1,89 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if message.content is string %}
+        {%- set content = message.content %}
+    {%- else %}
+        {%- set content = '' %}
+    {%- endif %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in content %}
+                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is false %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/checkpoint-9000/chat_template.jinja b/checkpoint-9000/chat_template.jinja
new file mode 100644
index 0000000..01be9b3
--- /dev/null
+++ b/checkpoint-9000/chat_template.jinja
@@ -0,0 +1,89 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if message.content is string %}
+        {%- set content = message.content %}
+    {%- else %}
+        {%- set content = '' %}
+    {%- endif %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in content %}
+                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is false %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/checkpoint-9000/config.json b/checkpoint-9000/config.json
new file mode 100644
index 0000000..1d219d6
--- /dev/null
+++ b/checkpoint-9000/config.json
@@ -0,0 +1,63 @@
+{
+  "architectures": [
+    "Qwen3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": null,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 40960,
+  "max_window_layers": 28,
+  "model_type": "qwen3",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 1000000,
+    "rope_type": "default"
+  },
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.7.0",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
diff --git a/checkpoint-9000/generation_config.json b/checkpoint-9000/generation_config.json
new file mode 100644
index 0000000..5ec133d
--- /dev/null
+++ b/checkpoint-9000/generation_config.json
@@ -0,0 +1,12 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "temperature": 0.6,
+  "top_k": 20,
+  "top_p": 0.95,
+  "transformers_version": "5.7.0"
+}
diff --git a/checkpoint-9000/model.safetensors b/checkpoint-9000/model.safetensors
new file mode 100644
index 0000000..e1de158
--- /dev/null
+++ b/checkpoint-9000/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:948f17ca34e5c7b09cfc8beaae1af7daae96e9fe08030b7f7bb59911421b1b66
+size 3441185608
diff --git a/checkpoint-9000/optimizer.pt b/checkpoint-9000/optimizer.pt
new file mode 100644
index 0000000..f09a14e
--- /dev/null
+++ b/checkpoint-9000/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:046c03d7bbdef1eca916da0e549f200a5b0083cc788ce297f7d93d1d01534b65
+size 6882572207
diff --git a/checkpoint-9000/rng_state.pth b/checkpoint-9000/rng_state.pth
new file mode 100644
index 0000000..6928c3e
--- /dev/null
+++ b/checkpoint-9000/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e51a42ec5ffac256f3cd0ee0ccd7a6d3befe58914ad400c0cc707bc4a5ee283
+size 14645
diff --git a/checkpoint-9000/scheduler.pt b/checkpoint-9000/scheduler.pt
new file mode 100644
index 0000000..36f5fde
--- /dev/null
+++ b/checkpoint-9000/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67bca42526be3b0972feec787047fe48501e9f96a69302aee50aa9e3ec5c5ace
+size 1465
diff --git a/checkpoint-9000/tokenizer.json b/checkpoint-9000/tokenizer.json
new file mode 100644
index 0000000..c7afbed
--- /dev/null
+++ b/checkpoint-9000/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
+size 11422650
diff --git a/checkpoint-9000/tokenizer_config.json b/checkpoint-9000/tokenizer_config.json
new file mode 100644
index 0000000..770e41d
--- /dev/null
+++ b/checkpoint-9000/tokenizer_config.json
@@ -0,0 +1,30 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "local_files_only": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}
diff --git a/checkpoint-9000/trainer_state.json b/checkpoint-9000/trainer_state.json
new file mode 100644
index 0000000..72270af
--- /dev/null
+++ b/checkpoint-9000/trainer_state.json
@@ -0,0 +1,4732 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.964824120603015,
+  "eval_steps": 500,
+  "global_step": 9000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 0.3966331691481173,
+      "epoch": 0.0021440536013400337,
+      "grad_norm": 129.0,
+      "learning_rate": 1.3571428571428572e-06,
+      "loss": 3.2754512786865235,
+      "mean_token_accuracy": 0.7025479640811682,
+      "num_tokens": 100865.0,
+      "step": 20
+    },
+    {
+      "entropy": 0.39382178746163843,
+      "epoch": 0.004288107202680067,
+      "grad_norm": 72.5,
+      "learning_rate": 2.785714285714286e-06,
+      "loss": 2.5979455947875976,
+      "mean_token_accuracy": 0.7279976762831211,
+      "num_tokens": 209237.0,
+      "step": 40
+    },
+    {
+      "entropy": 0.49158020988106726,
+      "epoch": 0.006432160804020101,
+      "grad_norm": 31.0,
+      "learning_rate": 4.2142857142857145e-06,
+      "loss": 1.34407320022583,
+      "mean_token_accuracy": 0.7954290494322777,
+      "num_tokens": 322722.0,
+      "step": 60
+    },
+    {
+      "entropy": 0.37321625389158725,
+      "epoch": 0.008576214405360135,
+      "grad_norm": 9.6875,
+      "learning_rate": 5.6428571428571435e-06,
+      "loss": 0.308458948135376,
+      "mean_token_accuracy": 0.9197555683553219,
+      "num_tokens": 426557.0,
+      "step": 80
+    },
+    {
+      "entropy": 0.09664252077927812,
+      "epoch": 0.010720268006700168,
+      "grad_norm": 4.25,
+      "learning_rate": 7.0714285714285726e-06,
+      "loss": 0.08551757931709289,
+      "mean_token_accuracy": 0.9772517420351505,
+      "num_tokens": 538043.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.06678911932976916,
+      "epoch": 0.012864321608040201,
+      "grad_norm": 6.59375,
+      "learning_rate": 8.5e-06,
+      "loss": 0.05725666880607605,
+      "mean_token_accuracy": 0.978511106595397,
+      "num_tokens": 651143.0,
+      "step": 120
+    },
+    {
+      "entropy": 0.05399178955703974,
+      "epoch": 0.015008375209380235,
+      "grad_norm": 1.8515625,
+      "learning_rate": 9.92857142857143e-06,
+      "loss": 0.0476167619228363,
+      "mean_token_accuracy": 0.9818631950765848,
+      "num_tokens": 759324.0,
+      "step": 140
+    },
+    {
+      "entropy": 0.050143458176171406,
+      "epoch": 0.01715242881072027,
+      "grad_norm": 2.65625,
+      "learning_rate": 1.1357142857142858e-05,
+      "loss": 0.048712074756622314,
+      "mean_token_accuracy": 0.9795998100191354,
+      "num_tokens": 854771.0,
+      "step": 160
+    },
+    {
+      "entropy": 0.04800442266860046,
+      "epoch": 0.019296482412060303,
+      "grad_norm": 2.53125,
+      "learning_rate": 1.2785714285714286e-05,
+      "loss": 0.046506857872009276,
+      "mean_token_accuracy": 0.9819129385054112,
+      "num_tokens": 965263.0,
+      "step": 180
+    },
+    {
+      "entropy": 0.04474863488285337,
+      "epoch": 0.021440536013400336,
+      "grad_norm": 2.46875,
+      "learning_rate": 1.4214285714285716e-05,
+      "loss": 0.04230240285396576,
+      "mean_token_accuracy": 0.9835243009030818,
+      "num_tokens": 1067106.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.04506044688168913,
+      "epoch": 0.02358458961474037,
+      "grad_norm": 4.59375,
+      "learning_rate": 1.5642857142857143e-05,
+      "loss": 0.04433030188083649,
+      "mean_token_accuracy": 0.9820781596004963,
+      "num_tokens": 1181133.0,
+      "step": 220
+    },
+    {
+      "entropy": 0.04305547107360326,
+      "epoch": 0.025728643216080403,
+      "grad_norm": 1.7890625,
+      "learning_rate": 1.7071428571428573e-05,
+      "loss": 0.04099105894565582,
+      "mean_token_accuracy": 0.9831817526370287,
+      "num_tokens": 1301484.0,
+      "step": 240
+    },
+    {
+      "entropy": 0.04266677545965649,
+      "epoch": 0.027872696817420436,
+      "grad_norm": 2.34375,
+      "learning_rate": 1.8500000000000002e-05,
+      "loss": 0.04160386621952057,
+      "mean_token_accuracy": 0.9813323471695184,
+      "num_tokens": 1403952.0,
+      "step": 260
+    },
+    {
+      "entropy": 0.04101881538517773,
+      "epoch": 0.03001675041876047,
+      "grad_norm": 2.59375,
+      "learning_rate": 1.992857142857143e-05,
+      "loss": 0.042915910482406616,
+      "mean_token_accuracy": 0.9821092572063208,
+      "num_tokens": 1517588.0,
+      "step": 280
+    },
+    {
+      "entropy": 0.04049238133156905,
+      "epoch": 0.032160804020100506,
+      "grad_norm": 2.703125,
+      "learning_rate": 1.9999782442429437e-05,
+      "loss": 0.04166347980499267,
+      "mean_token_accuracy": 0.9828211534768343,
+      "num_tokens": 1629657.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.04230240093893371,
+      "epoch": 0.03430485762144054,
+      "grad_norm": 1.2421875,
+      "learning_rate": 1.9999083376151357e-05,
+      "loss": 0.041567260026931764,
+      "mean_token_accuracy": 0.9822345357388258,
+      "num_tokens": 1721145.0,
+      "step": 320
+    },
+    {
+      "entropy": 0.03799658405041555,
+      "epoch": 0.03644891122278057,
+      "grad_norm": 3.3125,
+      "learning_rate": 1.9997902232214542e-05,
+      "loss": 0.0413459450006485,
+      "mean_token_accuracy": 0.9830254316329956,
+      "num_tokens": 1820492.0,
+      "step": 340
+    },
+    {
+      "entropy": 0.041703186882659796,
+      "epoch": 0.038592964824120606,
+      "grad_norm": 2.65625,
+      "learning_rate": 1.999623906756452e-05,
+      "loss": 0.040859296917915344,
+      "mean_token_accuracy": 0.976021521165967,
+      "num_tokens": 1931348.0,
+      "step": 360
+    },
+    {
+      "entropy": 0.04301113227265887,
+      "epoch": 0.04073701842546064,
+      "grad_norm": 1.7734375,
+      "learning_rate": 1.9994093962386102e-05,
+      "loss": 0.04014136493206024,
+      "mean_token_accuracy": 0.9837368253618479,
+      "num_tokens": 2047917.0,
+      "step": 380
+    },
+    {
+      "entropy": 0.041214162003598176,
+      "epoch": 0.04288107202680067,
+      "grad_norm": 1.5078125,
+      "learning_rate": 1.9991467020099492e-05,
+      "loss": 0.04444628953933716,
+      "mean_token_accuracy": 0.9809734560549259,
+      "num_tokens": 2149033.0,
+      "step": 400
+    },
+    {
+      "entropy": 0.04038102811900899,
+      "epoch": 0.045025125628140705,
+      "grad_norm": 1.6796875,
+      "learning_rate": 1.998835836735531e-05,
+      "loss": 0.0405484139919281,
+      "mean_token_accuracy": 0.982579792290926,
+      "num_tokens": 2248970.0,
+      "step": 420
+    },
+    {
+      "entropy": 0.039991069887764755,
+      "epoch": 0.04716917922948074,
+      "grad_norm": 1.09375,
+      "learning_rate": 1.9984768154028508e-05,
+      "loss": 0.03874191641807556,
+      "mean_token_accuracy": 0.9835739009082317,
+      "num_tokens": 2350681.0,
+      "step": 440
+    },
+    {
+      "entropy": 0.03837411678614444,
+      "epoch": 0.04931323283082077,
+      "grad_norm": 1.4296875,
+      "learning_rate": 1.9980696553211108e-05,
+      "loss": 0.04068012237548828,
+      "mean_token_accuracy": 0.9836468525230885,
+      "num_tokens": 2455656.0,
+      "step": 460
+    },
+    {
+      "entropy": 0.03925063398201019,
+      "epoch": 0.051457286432160805,
+      "grad_norm": 2.140625,
+      "learning_rate": 1.9976143761203893e-05,
+      "loss": 0.04098735451698303,
+      "mean_token_accuracy": 0.9809495396912098,
+      "num_tokens": 2559532.0,
+      "step": 480
+    },
+    {
+      "entropy": 0.03997599564900156,
+      "epoch": 0.05360134003350084,
+      "grad_norm": 1.546875,
+      "learning_rate": 1.99711099975069e-05,
+      "loss": 0.03899767398834229,
+      "mean_token_accuracy": 0.9831298362463713,
+      "num_tokens": 2654243.0,
+      "step": 500
+    },
+    {
+      "epoch": 0.05360134003350084,
+      "eval_entropy": 0.038057172449305655,
+      "eval_loss": 0.03832634538412094,
+      "eval_mean_token_accuracy": 0.9846738157272339,
+      "eval_num_tokens": 2654243.0,
+      "eval_runtime": 20.3077,
+      "eval_samples_per_second": 36.932,
+      "eval_steps_per_second": 18.466,
+      "step": 500
+    },
+    {
+      "entropy": 0.040187044665799476,
+      "epoch": 0.05574539363484087,
+      "grad_norm": 3.796875,
+      "learning_rate": 1.9965595504808886e-05,
+      "loss": 0.03858818709850311,
+      "mean_token_accuracy": 0.9832367639988661,
+      "num_tokens": 2751255.0,
+      "step": 520
+    },
+    {
+      "entropy": 0.035708320523553994,
+      "epoch": 0.057889447236180905,
+      "grad_norm": 1.8515625,
+      "learning_rate": 1.9959600548975593e-05,
+      "loss": 0.035676625370979306,
+      "mean_token_accuracy": 0.9861446134746075,
+      "num_tokens": 2849554.0,
+      "step": 540
+    },
+    {
+      "entropy": 0.036129794413864144,
+      "epoch": 0.06003350083752094,
+      "grad_norm": 1.3828125,
+      "learning_rate": 1.9953125419036935e-05,
+      "loss": 0.03737228214740753,
+      "mean_token_accuracy": 0.9838687002658844,
+      "num_tokens": 2948412.0,
+      "step": 560
+    },
+    {
+      "entropy": 0.036765301373088734,
+      "epoch": 0.06217755443886097,
+      "grad_norm": 1.6640625,
+      "learning_rate": 1.994617042717308e-05,
+      "loss": 0.03550122678279877,
+      "mean_token_accuracy": 0.9847423356026411,
+      "num_tokens": 3048326.0,
+      "step": 580
+    },
+    {
+      "entropy": 0.03411570826137904,
+      "epoch": 0.06432160804020101,
+      "grad_norm": 2.40625,
+      "learning_rate": 1.9938735908699384e-05,
+      "loss": 0.036294499039649965,
+      "mean_token_accuracy": 0.9855027917772532,
+      "num_tokens": 3161774.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.0356013277938473,
+      "epoch": 0.06646566164154104,
+      "grad_norm": 1.84375,
+      "learning_rate": 1.993082222205023e-05,
+      "loss": 0.034555235505104066,
+      "mean_token_accuracy": 0.9851896297186613,
+      "num_tokens": 3273915.0,
+      "step": 620
+    },
+    {
+      "entropy": 0.0345983263629023,
+      "epoch": 0.06860971524288108,
+      "grad_norm": 2.34375,
+      "learning_rate": 1.9922429748761733e-05,
+      "loss": 0.03647570013999939,
+      "mean_token_accuracy": 0.9840687446296215,
+      "num_tokens": 3381550.0,
+      "step": 640
+    },
+    {
+      "entropy": 0.03738872102476307,
+      "epoch": 0.0707537688442211,
+      "grad_norm": 1.7265625,
+      "learning_rate": 1.9913558893453382e-05,
+      "loss": 0.036225003004074094,
+      "mean_token_accuracy": 0.9841504596173763,
+      "num_tokens": 3484439.0,
+      "step": 660
+    },
+    {
+      "entropy": 0.037293507948561455,
+      "epoch": 0.07289782244556114,
+      "grad_norm": 2.0625,
+      "learning_rate": 1.9904210083808485e-05,
+      "loss": 0.03906684815883636,
+      "mean_token_accuracy": 0.9821466624736785,
+      "num_tokens": 3588594.0,
+      "step": 680
+    },
+    {
+      "entropy": 0.03572316213103477,
+      "epoch": 0.07504187604690117,
+      "grad_norm": 1.3671875,
+      "learning_rate": 1.9894383770553594e-05,
+      "loss": 0.03395065367221832,
+      "mean_token_accuracy": 0.9846713658422231,
+      "num_tokens": 3690877.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.0373659498654888,
+      "epoch": 0.07718592964824121,
+      "grad_norm": 1.1328125,
+      "learning_rate": 1.988408042743673e-05,
+      "loss": 0.03977937400341034,
+      "mean_token_accuracy": 0.982515212148428,
+      "num_tokens": 3802248.0,
+      "step": 720
+    },
+    {
+      "entropy": 0.03811679463542532,
+      "epoch": 0.07932998324958124,
+      "grad_norm": 3.171875,
+      "learning_rate": 1.9873300551204588e-05,
+      "loss": 0.03767184019088745,
+      "mean_token_accuracy": 0.9842769052833319,
+      "num_tokens": 3904946.0,
+      "step": 740
+    },
+    {
+      "entropy": 0.03399244465199445,
+      "epoch": 0.08147403685092128,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.9862044661578554e-05,
+      "loss": 0.033424127101898196,
+      "mean_token_accuracy": 0.9856430813670158,
+      "num_tokens": 4011765.0,
+      "step": 760
+    },
+    {
+      "entropy": 0.03507241023471579,
+      "epoch": 0.0836180904522613,
+      "grad_norm": 2.328125,
+      "learning_rate": 1.985031330122966e-05,
+      "loss": 0.034555363655090335,
+      "mean_token_accuracy": 0.9848727196455002,
+      "num_tokens": 4119751.0,
+      "step": 780
+    },
+    {
+      "entropy": 0.03295613579248311,
+      "epoch": 0.08576214405360134,
+      "grad_norm": 2.015625,
+      "learning_rate": 1.9838107035752422e-05,
+      "loss": 0.03498654961585999,
+      "mean_token_accuracy": 0.9861093372106552,
+      "num_tokens": 4234054.0,
+      "step": 800
+    },
+    {
+      "entropy": 0.03774066482146736,
+      "epoch": 0.08790619765494137,
+      "grad_norm": 1.453125,
+      "learning_rate": 1.9825426453637578e-05,
+      "loss": 0.03699869811534882,
+      "mean_token_accuracy": 0.9836779370903969,
+      "num_tokens": 4345322.0,
+      "step": 820
+    },
+    {
+      "entropy": 0.03479030575981597,
+      "epoch": 0.09005025125628141,
+      "grad_norm": 1.8984375,
+      "learning_rate": 1.9812272166243694e-05,
+      "loss": 0.03539972007274628,
+      "mean_token_accuracy": 0.98408287987113,
+      "num_tokens": 4439690.0,
+      "step": 840
+    },
+    {
+      "entropy": 0.03287471420262591,
+      "epoch": 0.09219430485762144,
+      "grad_norm": 1.9375,
+      "learning_rate": 1.979864480776771e-05,
+      "loss": 0.033072105050086974,
+      "mean_token_accuracy": 0.9868043974041939,
+      "num_tokens": 4547696.0,
+      "step": 860
+    },
+    {
+      "entropy": 0.03525375899625942,
+      "epoch": 0.09433835845896148,
+      "grad_norm": 1.3984375,
+      "learning_rate": 1.9784545035214358e-05,
+      "loss": 0.034766626358032224,
+      "mean_token_accuracy": 0.985904373601079,
+      "num_tokens": 4656576.0,
+      "step": 880
+    },
+    {
+      "entropy": 0.03227176534710452,
+      "epoch": 0.0964824120603015,
+      "grad_norm": 1.46875,
+      "learning_rate": 1.976997352836448e-05,
+      "loss": 0.032944831252098086,
+      "mean_token_accuracy": 0.9853209290653467,
+      "num_tokens": 4763714.0,
+      "step": 900
+    },
+    {
+      "entropy": 0.03179633504041703,
+      "epoch": 0.09862646566164154,
+      "grad_norm": 1.5,
+      "learning_rate": 1.975493098974226e-05,
+      "loss": 0.033859679102897645,
+      "mean_token_accuracy": 0.9853155620396137,
+      "num_tokens": 4862673.0,
+      "step": 920
+    },
+    {
+      "entropy": 0.0342723391513573,
+      "epoch": 0.10077051926298157,
+      "grad_norm": 1.703125,
+      "learning_rate": 1.9739418144581357e-05,
+      "loss": 0.035592520236968996,
+      "mean_token_accuracy": 0.984546373412013,
+      "num_tokens": 4972462.0,
+      "step": 940
+    },
+    {
+      "entropy": 0.031624948812532236,
+      "epoch": 0.10291457286432161,
+      "grad_norm": 1.875,
+      "learning_rate": 1.9723435740789937e-05,
+      "loss": 0.03236276209354401,
+      "mean_token_accuracy": 0.9860632043331862,
+      "num_tokens": 5076752.0,
+      "step": 960
+    },
+    {
+      "entropy": 0.03294982312654611,
+      "epoch": 0.10505862646566164,
+      "grad_norm": 1.9140625,
+      "learning_rate": 1.97069845489146e-05,
+      "loss": 0.03309372365474701,
+      "mean_token_accuracy": 0.9859170567244291,
+      "num_tokens": 5194841.0,
+      "step": 980
+    },
+    {
+      "entropy": 0.03638018769561313,
+      "epoch": 0.10720268006700168,
+      "grad_norm": 1.3671875,
+      "learning_rate": 1.969006536210327e-05,
+      "loss": 0.036982369422912595,
+      "mean_token_accuracy": 0.9845436029136181,
+      "num_tokens": 5295536.0,
+      "step": 1000
+    },
+    {
+      "epoch": 0.10720268006700168,
+      "eval_entropy": 0.03662689963728189,
+      "eval_loss": 0.033028148114681244,
+      "eval_mean_token_accuracy": 0.9869652398427328,
+      "eval_num_tokens": 5295536.0,
+      "eval_runtime": 20.3094,
+      "eval_samples_per_second": 36.929,
+      "eval_steps_per_second": 18.464,
+      "step": 1000
+    },
+    {
+      "entropy": 0.035084289042788444,
+      "epoch": 0.1093467336683417,
+      "grad_norm": 1.4609375,
+      "learning_rate": 1.96726789960669e-05,
+      "loss": 0.03461971580982208,
+      "mean_token_accuracy": 0.9852123014628887,
+      "num_tokens": 5398244.0,
+      "step": 1020
+    },
+    {
+      "entropy": 0.03179455297176901,
+      "epoch": 0.11149078726968174,
+      "grad_norm": 1.3515625,
+      "learning_rate": 1.965482628904019e-05,
+      "loss": 0.03005821704864502,
+      "mean_token_accuracy": 0.9795859795063734,
+      "num_tokens": 5513313.0,
+      "step": 1040
+    },
+    {
+      "entropy": 0.028698108330354445,
+      "epoch": 0.11363484087102177,
+      "grad_norm": 3.4375,
+      "learning_rate": 1.963650810174116e-05,
+      "loss": 0.03329857885837555,
+      "mean_token_accuracy": 0.9862619318068028,
+      "num_tokens": 5612778.0,
+      "step": 1060
+    },
+    {
+      "entropy": 0.03370735585340299,
+      "epoch": 0.11577889447236181,
+      "grad_norm": 1.84375,
+      "learning_rate": 1.9617725317329644e-05,
+      "loss": 0.03324403166770935,
+      "mean_token_accuracy": 0.985461600497365,
+      "num_tokens": 5719033.0,
+      "step": 1080
+    },
+    {
+      "entropy": 0.035593700742174404,
+      "epoch": 0.11792294807370184,
+      "grad_norm": 1.8671875,
+      "learning_rate": 1.9598478841364727e-05,
+      "loss": 0.034142997860908506,
+      "mean_token_accuracy": 0.985479149222374,
+      "num_tokens": 5833790.0,
+      "step": 1100
+    },
+    {
+      "entropy": 0.029760706042361564,
+      "epoch": 0.12006700167504188,
+      "grad_norm": 1.234375,
+      "learning_rate": 1.957876960176106e-05,
+      "loss": 0.03303615748882294,
+      "mean_token_accuracy": 0.9864209692925214,
+      "num_tokens": 5942042.0,
+      "step": 1120
+    },
+    {
+      "entropy": 0.03119294007847202,
+      "epoch": 0.12221105527638192,
+      "grad_norm": 2.234375,
+      "learning_rate": 1.955859854874416e-05,
+      "loss": 0.03496485948562622,
+      "mean_token_accuracy": 0.9869707945734263,
+      "num_tokens": 6052169.0,
+      "step": 1140
+    },
+    {
+      "entropy": 0.032620664288697296,
+      "epoch": 0.12435510887772194,
+      "grad_norm": 2.1875,
+      "learning_rate": 1.9537966654804568e-05,
+      "loss": 0.032474163174629214,
+      "mean_token_accuracy": 0.985412934049964,
+      "num_tokens": 6160024.0,
+      "step": 1160
+    },
+    {
+      "entropy": 0.029878590396401707,
+      "epoch": 0.12649916247906198,
+      "grad_norm": 1.3203125,
+      "learning_rate": 1.9516874914650972e-05,
+      "loss": 0.028646692633628845,
+      "mean_token_accuracy": 0.9869820442050695,
+      "num_tokens": 6264476.0,
+      "step": 1180
+    },
+    {
+      "entropy": 0.03366943938890472,
+      "epoch": 0.12864321608040202,
+      "grad_norm": 2.390625,
+      "learning_rate": 1.9495324345162255e-05,
+      "loss": 0.03512271046638489,
+      "mean_token_accuracy": 0.9845213159918785,
+      "num_tokens": 6377469.0,
+      "step": 1200
+    },
+    {
+      "entropy": 0.03266510410194314,
+      "epoch": 0.13078726968174204,
+      "grad_norm": 1.625,
+      "learning_rate": 1.947331598533846e-05,
+      "loss": 0.03071320950984955,
+      "mean_token_accuracy": 0.9855085987597704,
+      "num_tokens": 6489484.0,
+      "step": 1220
+    },
+    {
+      "entropy": 0.02746599452148075,
+      "epoch": 0.13293132328308208,
+      "grad_norm": 2.03125,
+      "learning_rate": 1.9450850896250705e-05,
+      "loss": 0.029182371497154237,
+      "mean_token_accuracy": 0.9887533523142338,
+      "num_tokens": 6594306.0,
+      "step": 1240
+    },
+    {
+      "entropy": 0.029792005700437586,
+      "epoch": 0.13507537688442212,
+      "grad_norm": 1.8359375,
+      "learning_rate": 1.9427930160990026e-05,
+      "loss": 0.03335131406784057,
+      "mean_token_accuracy": 0.9871865373104811,
+      "num_tokens": 6701348.0,
+      "step": 1260
+    },
+    {
+      "entropy": 0.03408827395323897,
+      "epoch": 0.13721943048576216,
+      "grad_norm": 1.40625,
+      "learning_rate": 1.9404554884615153e-05,
+      "loss": 0.033723458647727966,
+      "mean_token_accuracy": 0.9851170387119055,
+      "num_tokens": 6812895.0,
+      "step": 1280
+    },
+    {
+      "entropy": 0.03265384387705126,
+      "epoch": 0.13936348408710217,
+      "grad_norm": 2.125,
+      "learning_rate": 1.9380726194099236e-05,
+      "loss": 0.032703128457069394,
+      "mean_token_accuracy": 0.9860071249306201,
+      "num_tokens": 6914997.0,
+      "step": 1300
+    },
+    {
+      "entropy": 0.030018781982835207,
+      "epoch": 0.1415075376884422,
+      "grad_norm": 1.9765625,
+      "learning_rate": 1.9356445238275518e-05,
+      "loss": 0.029086670279502867,
+      "mean_token_accuracy": 0.9875706855207682,
+      "num_tokens": 7017334.0,
+      "step": 1320
+    },
+    {
+      "entropy": 0.030701546505588338,
+      "epoch": 0.14365159128978225,
+      "grad_norm": 2.4375,
+      "learning_rate": 1.933171318778193e-05,
+      "loss": 0.03185203075408936,
+      "mean_token_accuracy": 0.9859924230724573,
+      "num_tokens": 7124252.0,
+      "step": 1340
+    },
+    {
+      "entropy": 0.032219707980402745,
+      "epoch": 0.1457956448911223,
+      "grad_norm": 1.359375,
+      "learning_rate": 1.9306531235004674e-05,
+      "loss": 0.03223040103912354,
+      "mean_token_accuracy": 0.9870269387960434,
+      "num_tokens": 7234001.0,
+      "step": 1360
+    },
+    {
+      "entropy": 0.029969332017935814,
+      "epoch": 0.1479396984924623,
+      "grad_norm": 1.578125,
+      "learning_rate": 1.928090059402072e-05,
+      "loss": 0.030724048614501953,
+      "mean_token_accuracy": 0.9877043284475804,
+      "num_tokens": 7340453.0,
+      "step": 1380
+    },
+    {
+      "entropy": 0.031696403899695726,
+      "epoch": 0.15008375209380234,
+      "grad_norm": 1.25,
+      "learning_rate": 1.9254822500539268e-05,
+      "loss": 0.032300516963005066,
+      "mean_token_accuracy": 0.9876017738133669,
+      "num_tokens": 7453174.0,
+      "step": 1400
+    },
+    {
+      "entropy": 0.0315240634514339,
+      "epoch": 0.15222780569514238,
+      "grad_norm": 1.328125,
+      "learning_rate": 1.9228298211842193e-05,
+      "loss": 0.033043709397315976,
+      "mean_token_accuracy": 0.9866235084831715,
+      "num_tokens": 7555849.0,
+      "step": 1420
+    },
+    {
+      "entropy": 0.031768915360407844,
+      "epoch": 0.15437185929648242,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.9201329006723413e-05,
+      "loss": 0.027889394760131837,
+      "mean_token_accuracy": 0.9881493493914604,
+      "num_tokens": 7665192.0,
+      "step": 1440
+    },
+    {
+      "entropy": 0.02793313514721376,
+      "epoch": 0.15651591289782243,
+      "grad_norm": 1.6015625,
+      "learning_rate": 1.917391618542724e-05,
+      "loss": 0.034602677822113036,
+      "mean_token_accuracy": 0.9850024841725826,
+      "num_tokens": 7766528.0,
+      "step": 1460
+    },
+    {
+      "entropy": 0.033910460214247,
+      "epoch": 0.15865996649916247,
+      "grad_norm": 2.375,
+      "learning_rate": 1.914606106958568e-05,
+      "loss": 0.03377147018909454,
+      "mean_token_accuracy": 0.9864576529711485,
+      "num_tokens": 7873249.0,
+      "step": 1480
+    },
+    {
+      "entropy": 0.03134421622089576,
+      "epoch": 0.16080402010050251,
+      "grad_norm": 1.6328125,
+      "learning_rate": 1.911776500215474e-05,
+      "loss": 0.031429946422576904,
+      "mean_token_accuracy": 0.9866740386933088,
+      "num_tokens": 7966983.0,
+      "step": 1500
+    },
+    {
+      "epoch": 0.16080402010050251,
+      "eval_entropy": 0.031583791558475544,
+      "eval_loss": 0.029090270400047302,
+      "eval_mean_token_accuracy": 0.9873407395680746,
+      "eval_num_tokens": 7966983.0,
+      "eval_runtime": 20.2978,
+      "eval_samples_per_second": 36.95,
+      "eval_steps_per_second": 18.475,
+      "step": 1500
+    },
+    {
+      "entropy": 0.03011373901317711,
+      "epoch": 0.16294807370184256,
+      "grad_norm": 1.96875,
+      "learning_rate": 1.9089029347349664e-05,
+      "loss": 0.029482910037040712,
+      "mean_token_accuracy": 0.9875899031758308,
+      "num_tokens": 8076659.0,
+      "step": 1520
+    },
+    {
+      "entropy": 0.024959181708436518,
+      "epoch": 0.16509212730318257,
+      "grad_norm": 2.875,
+      "learning_rate": 1.9059855490579156e-05,
+      "loss": 0.02683677673339844,
+      "mean_token_accuracy": 0.989487374201417,
+      "num_tokens": 8175314.0,
+      "step": 1540
+    },
+    {
+      "entropy": 0.027774957335896033,
+      "epoch": 0.1672361809045226,
+      "grad_norm": 2.09375,
+      "learning_rate": 1.90302448383786e-05,
+      "loss": 0.03151543140411377,
+      "mean_token_accuracy": 0.9872768469154835,
+      "num_tokens": 8290905.0,
+      "step": 1560
+    },
+    {
+      "entropy": 0.03308429398821318,
+      "epoch": 0.16938023450586265,
+      "grad_norm": 1.203125,
+      "learning_rate": 1.9000198818342243e-05,
+      "loss": 0.03073907792568207,
+      "mean_token_accuracy": 0.9868967544287444,
+      "num_tokens": 8397808.0,
+      "step": 1580
+    },
+    {
+      "entropy": 0.030384358397623146,
+      "epoch": 0.1715242881072027,
+      "grad_norm": 1.9296875,
+      "learning_rate": 1.8969718879054365e-05,
+      "loss": 0.034263554215431216,
+      "mean_token_accuracy": 0.98636131323874,
+      "num_tokens": 8504583.0,
+      "step": 1600
+    },
+    {
+      "entropy": 0.02909080486315361,
+      "epoch": 0.1736683417085427,
+      "grad_norm": 2.078125,
+      "learning_rate": 1.893880649001944e-05,
+      "loss": 0.029698848724365234,
+      "mean_token_accuracy": 0.9875615444034338,
+      "num_tokens": 8608889.0,
+      "step": 1620
+    },
+    {
+      "entropy": 0.03127812161601469,
+      "epoch": 0.17581239530988274,
+      "grad_norm": 2.3125,
+      "learning_rate": 1.890746314159129e-05,
+      "loss": 0.031479662656784056,
+      "mean_token_accuracy": 0.9861587047576904,
+      "num_tokens": 8720168.0,
+      "step": 1640
+    },
+    {
+      "entropy": 0.028632436878615407,
+      "epoch": 0.17795644891122278,
+      "grad_norm": 2.96875,
+      "learning_rate": 1.8875690344901237e-05,
+      "loss": 0.02911217212677002,
+      "mean_token_accuracy": 0.9880903176963329,
+      "num_tokens": 8822057.0,
+      "step": 1660
+    },
+    {
+      "entropy": 0.02712122074663057,
+      "epoch": 0.18010050251256282,
+      "grad_norm": 5.09375,
+      "learning_rate": 1.884348963178523e-05,
+      "loss": 0.02785937786102295,
+      "mean_token_accuracy": 0.9886827927082777,
+      "num_tokens": 8926336.0,
+      "step": 1680
+    },
+    {
+      "entropy": 0.025899743908667006,
+      "epoch": 0.18224455611390286,
+      "grad_norm": 1.265625,
+      "learning_rate": 1.881086255471001e-05,
+      "loss": 0.026762637495994567,
+      "mean_token_accuracy": 0.9886533226817846,
+      "num_tokens": 9023646.0,
+      "step": 1700
+    },
+    {
+      "entropy": 0.028882844069084968,
+      "epoch": 0.18438860971524287,
+      "grad_norm": 1.5390625,
+      "learning_rate": 1.8777810686698266e-05,
+      "loss": 0.029647400975227355,
+      "mean_token_accuracy": 0.9875607810914516,
+      "num_tokens": 9127152.0,
+      "step": 1720
+    },
+    {
+      "entropy": 0.03044446731946664,
+      "epoch": 0.18653266331658291,
+      "grad_norm": 1.2265625,
+      "learning_rate": 1.8744335621252766e-05,
+      "loss": 0.031170135736465453,
+      "mean_token_accuracy": 0.9866547036916018,
+      "num_tokens": 9234279.0,
+      "step": 1740
+    },
+    {
+      "entropy": 0.029704924078851037,
+      "epoch": 0.18867671691792295,
+      "grad_norm": 1.546875,
+      "learning_rate": 1.8710438972279572e-05,
+      "loss": 0.030961766839027405,
+      "mean_token_accuracy": 0.9872025232762098,
+      "num_tokens": 9348546.0,
+      "step": 1760
+    },
+    {
+      "entropy": 0.026529241763273603,
+      "epoch": 0.190820770519263,
+      "grad_norm": 2.828125,
+      "learning_rate": 1.867612237401019e-05,
+      "loss": 0.026679041981697082,
+      "mean_token_accuracy": 0.9885351285338402,
+      "num_tokens": 9455489.0,
+      "step": 1780
+    },
+    {
+      "entropy": 0.028892852607532403,
+      "epoch": 0.192964824120603,
+      "grad_norm": 4.03125,
+      "learning_rate": 1.8641387480922816e-05,
+      "loss": 0.03060615360736847,
+      "mean_token_accuracy": 0.9880913089960813,
+      "num_tokens": 9561715.0,
+      "step": 1800
+    },
+    {
+      "entropy": 0.03127149782612833,
+      "epoch": 0.19510887772194305,
+      "grad_norm": 1.4375,
+      "learning_rate": 1.860623596766254e-05,
+      "loss": 0.03185727298259735,
+      "mean_token_accuracy": 0.9862949691712857,
+      "num_tokens": 9671395.0,
+      "step": 1820
+    },
+    {
+      "entropy": 0.030052698707368108,
+      "epoch": 0.1972529313232831,
+      "grad_norm": 1.7265625,
+      "learning_rate": 1.8570669528960627e-05,
+      "loss": 0.030486562848091127,
+      "mean_token_accuracy": 0.9874879900366068,
+      "num_tokens": 9772135.0,
+      "step": 1840
+    },
+    {
+      "entropy": 0.028759777540108188,
+      "epoch": 0.19939698492462313,
+      "grad_norm": 1.5546875,
+      "learning_rate": 1.8534689879552806e-05,
+      "loss": 0.02927035093307495,
+      "mean_token_accuracy": 0.9883296392858029,
+      "num_tokens": 9875332.0,
+      "step": 1860
+    },
+    {
+      "entropy": 0.026194313674932346,
+      "epoch": 0.20154103852596314,
+      "grad_norm": 1.2421875,
+      "learning_rate": 1.8498298754096597e-05,
+      "loss": 0.027384281158447266,
+      "mean_token_accuracy": 0.9899293314665556,
+      "num_tokens": 9975866.0,
+      "step": 1880
+    },
+    {
+      "entropy": 0.030918022169134928,
+      "epoch": 0.20368509212730318,
+      "grad_norm": 1.7421875,
+      "learning_rate": 1.8461497907087672e-05,
+      "loss": 0.03030203580856323,
+      "mean_token_accuracy": 0.9866726931184531,
+      "num_tokens": 10091897.0,
+      "step": 1900
+    },
+    {
+      "entropy": 0.02552133451026748,
+      "epoch": 0.20582914572864322,
+      "grad_norm": 1.1875,
+      "learning_rate": 1.8424289112775284e-05,
+      "loss": 0.027020731568336488,
+      "mean_token_accuracy": 0.9879696272313595,
+      "num_tokens": 10189590.0,
+      "step": 1920
+    },
+    {
+      "entropy": 0.026746707828442594,
+      "epoch": 0.20797319932998326,
+      "grad_norm": 3.59375,
+      "learning_rate": 1.8386674165076706e-05,
+      "loss": 0.026832282543182373,
+      "mean_token_accuracy": 0.9871037002652884,
+      "num_tokens": 10290612.0,
+      "step": 1940
+    },
+    {
+      "entropy": 0.023915313509087356,
+      "epoch": 0.21011725293132327,
+      "grad_norm": 3.75,
+      "learning_rate": 1.8348654877490765e-05,
+      "loss": 0.03067324459552765,
+      "mean_token_accuracy": 0.9885836262255907,
+      "num_tokens": 10400962.0,
+      "step": 1960
+    },
+    {
+      "entropy": 0.027086517597672354,
+      "epoch": 0.2122613065326633,
+      "grad_norm": 2.078125,
+      "learning_rate": 1.831023308301039e-05,
+      "loss": 0.028256928920745848,
+      "mean_token_accuracy": 0.9885173581540585,
+      "num_tokens": 10506362.0,
+      "step": 1980
+    },
+    {
+      "entropy": 0.02741672217816813,
+      "epoch": 0.21440536013400335,
+      "grad_norm": 0.9140625,
+      "learning_rate": 1.827141063403425e-05,
+      "loss": 0.02612453103065491,
+      "mean_token_accuracy": 0.9899405423551798,
+      "num_tokens": 10614125.0,
+      "step": 2000
+    },
+    {
+      "epoch": 0.21440536013400335,
+      "eval_entropy": 0.024848208941170014,
+      "eval_loss": 0.026614099740982056,
+      "eval_mean_token_accuracy": 0.9895401794115702,
+      "eval_num_tokens": 10614125.0,
+      "eval_runtime": 20.3696,
+      "eval_samples_per_second": 36.82,
+      "eval_steps_per_second": 18.41,
+      "step": 2000
+    },
+    {
+      "entropy": 0.024184393659743363,
+      "epoch": 0.2165494137353434,
+      "grad_norm": 1.1015625,
+      "learning_rate": 1.8232189402277446e-05,
+      "loss": 0.029155582189559937,
+      "mean_token_accuracy": 0.9877722147852183,
+      "num_tokens": 10718545.0,
+      "step": 2020
+    },
+    {
+      "entropy": 0.026252732726061368,
+      "epoch": 0.2186934673366834,
+      "grad_norm": 1.8046875,
+      "learning_rate": 1.8192571278681253e-05,
+      "loss": 0.026098987460136412,
+      "mean_token_accuracy": 0.9895182501524686,
+      "num_tokens": 10826477.0,
+      "step": 2040
+    },
+    {
+      "entropy": 0.025545999853238754,
+      "epoch": 0.22083752093802345,
+      "grad_norm": 2.234375,
+      "learning_rate": 1.8152558173321987e-05,
+      "loss": 0.026094555854797363,
+      "mean_token_accuracy": 0.9890841003507376,
+      "num_tokens": 10925874.0,
+      "step": 2060
+    },
+    {
+      "entropy": 0.026384522972375635,
+      "epoch": 0.2229815745393635,
+      "grad_norm": 2.40625,
+      "learning_rate": 1.8112152015318894e-05,
+      "loss": 0.029251503944396972,
+      "mean_token_accuracy": 0.9880545783787966,
+      "num_tokens": 11033510.0,
+      "step": 2080
+    },
+    {
+      "entropy": 0.025210863275060547,
+      "epoch": 0.22512562814070353,
+      "grad_norm": 2.703125,
+      "learning_rate": 1.8071354752741137e-05,
+      "loss": 0.027124592661857606,
+      "mean_token_accuracy": 0.9892004739493132,
+      "num_tokens": 11137817.0,
+      "step": 2100
+    },
+    {
+      "entropy": 0.027089714014164202,
+      "epoch": 0.22726968174204354,
+      "grad_norm": 1.7578125,
+      "learning_rate": 1.8030168352513894e-05,
+      "loss": 0.026511949300765992,
+      "mean_token_accuracy": 0.9889535740017891,
+      "num_tokens": 11245185.0,
+      "step": 2120
+    },
+    {
+      "entropy": 0.024535768478563112,
+      "epoch": 0.22941373534338358,
+      "grad_norm": 4.625,
+      "learning_rate": 1.7988594800323515e-05,
+      "loss": 0.02658265233039856,
+      "mean_token_accuracy": 0.9905748564749957,
+      "num_tokens": 11349330.0,
+      "step": 2140
+    },
+    {
+      "entropy": 0.02754759546187415,
+      "epoch": 0.23155778894472362,
+      "grad_norm": 1.9609375,
+      "learning_rate": 1.7946636100521798e-05,
+      "loss": 0.027984172105789185,
+      "mean_token_accuracy": 0.9884784188121557,
+      "num_tokens": 11461079.0,
+      "step": 2160
+    },
+    {
+      "entropy": 0.025805097624288464,
+      "epoch": 0.23370184254606366,
+      "grad_norm": 2.078125,
+      "learning_rate": 1.7904294276029346e-05,
+      "loss": 0.026595547795295715,
+      "mean_token_accuracy": 0.9894516475498676,
+      "num_tokens": 11563866.0,
+      "step": 2180
+    },
+    {
+      "entropy": 0.026482580404353938,
+      "epoch": 0.23584589614740367,
+      "grad_norm": 2.765625,
+      "learning_rate": 1.7861571368238035e-05,
+      "loss": 0.02969781160354614,
+      "mean_token_accuracy": 0.9875592797994613,
+      "num_tokens": 11669471.0,
+      "step": 2200
+    },
+    {
+      "entropy": 0.028677671598052255,
+      "epoch": 0.2379899497487437,
+      "grad_norm": 2.21875,
+      "learning_rate": 1.7818469436912605e-05,
+      "loss": 0.02763218283653259,
+      "mean_token_accuracy": 0.9888417005538941,
+      "num_tokens": 11768609.0,
+      "step": 2220
+    },
+    {
+      "entropy": 0.026651295177725843,
+      "epoch": 0.24013400335008375,
+      "grad_norm": 1.859375,
+      "learning_rate": 1.7774990560091356e-05,
+      "loss": 0.02962457537651062,
+      "mean_token_accuracy": 0.9879184097051621,
+      "num_tokens": 11880854.0,
+      "step": 2240
+    },
+    {
+      "entropy": 0.02831397352310887,
+      "epoch": 0.2422780569514238,
+      "grad_norm": 1.8984375,
+      "learning_rate": 1.773113683398593e-05,
+      "loss": 0.0296058714389801,
+      "mean_token_accuracy": 0.9883440136909485,
+      "num_tokens": 11986546.0,
+      "step": 2260
+    },
+    {
+      "entropy": 0.0280173282616488,
+      "epoch": 0.24442211055276383,
+      "grad_norm": 1.9609375,
+      "learning_rate": 1.76869103728803e-05,
+      "loss": 0.026104235649108888,
+      "mean_token_accuracy": 0.9899998947978019,
+      "num_tokens": 12089099.0,
+      "step": 2280
+    },
+    {
+      "entropy": 0.027054303306613292,
+      "epoch": 0.24656616415410385,
+      "grad_norm": 1.7734375,
+      "learning_rate": 1.7642313309028798e-05,
+      "loss": 0.029236692190170287,
+      "mean_token_accuracy": 0.988573432713747,
+      "num_tokens": 12198358.0,
+      "step": 2300
+    },
+    {
+      "entropy": 0.025954095610904913,
+      "epoch": 0.24871021775544389,
+      "grad_norm": 2.53125,
+      "learning_rate": 1.7597347792553313e-05,
+      "loss": 0.02622365057468414,
+      "mean_token_accuracy": 0.9882505975663662,
+      "num_tokens": 12310140.0,
+      "step": 2320
+    },
+    {
+      "entropy": 0.02558089109697903,
+      "epoch": 0.2508542713567839,
+      "grad_norm": 2.5625,
+      "learning_rate": 1.7552015991339655e-05,
+      "loss": 0.025320738554000854,
+      "mean_token_accuracy": 0.9904545627534389,
+      "num_tokens": 12413395.0,
+      "step": 2340
+    },
+    {
+      "entropy": 0.02581585149782768,
+      "epoch": 0.25299832495812397,
+      "grad_norm": 1.59375,
+      "learning_rate": 1.7506320090933012e-05,
+      "loss": 0.02813691198825836,
+      "mean_token_accuracy": 0.9890731148421764,
+      "num_tokens": 12517491.0,
+      "step": 2360
+    },
+    {
+      "entropy": 0.029305921567720362,
+      "epoch": 0.255142378559464,
+      "grad_norm": 1.46875,
+      "learning_rate": 1.74602622944326e-05,
+      "loss": 0.030352601408958436,
+      "mean_token_accuracy": 0.9881349734961986,
+      "num_tokens": 12627025.0,
+      "step": 2380
+    },
+    {
+      "entropy": 0.026154012439656073,
+      "epoch": 0.25728643216080405,
+      "grad_norm": 2.03125,
+      "learning_rate": 1.7413844822385423e-05,
+      "loss": 0.0258339524269104,
+      "mean_token_accuracy": 0.9894504573196172,
+      "num_tokens": 12726236.0,
+      "step": 2400
+    },
+    {
+      "entropy": 0.024479785768926378,
+      "epoch": 0.25943048576214406,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.736706991267924e-05,
+      "loss": 0.027092832326889037,
+      "mean_token_accuracy": 0.9882367230951786,
+      "num_tokens": 12834614.0,
+      "step": 2420
+    },
+    {
+      "entropy": 0.02744944194200798,
+      "epoch": 0.26157453936348407,
+      "grad_norm": 2.734375,
+      "learning_rate": 1.7319939820434656e-05,
+      "loss": 0.026963105797767638,
+      "mean_token_accuracy": 0.9900378599762917,
+      "num_tokens": 12940452.0,
+      "step": 2440
+    },
+    {
+      "entropy": 0.02669895636208821,
+      "epoch": 0.26371859296482414,
+      "grad_norm": 6.9375,
+      "learning_rate": 1.72724568178964e-05,
+      "loss": 0.02737392783164978,
+      "mean_token_accuracy": 0.9884641785174608,
+      "num_tokens": 13038623.0,
+      "step": 2460
+    },
+    {
+      "entropy": 0.02623358482396725,
+      "epoch": 0.26586264656616415,
+      "grad_norm": 1.65625,
+      "learning_rate": 1.7224623194323785e-05,
+      "loss": 0.026823097467422487,
+      "mean_token_accuracy": 0.9890745796263218,
+      "num_tokens": 13145771.0,
+      "step": 2480
+    },
+    {
+      "entropy": 0.02654599146626424,
+      "epoch": 0.26800670016750416,
+      "grad_norm": 2.125,
+      "learning_rate": 1.717644125588032e-05,
+      "loss": 0.02797144651412964,
+      "mean_token_accuracy": 0.9888463616371155,
+      "num_tokens": 13258191.0,
+      "step": 2500
+    },
+    {
+      "epoch": 0.26800670016750416,
+      "eval_entropy": 0.02707736179015289,
+      "eval_loss": 0.024532999843358994,
+      "eval_mean_token_accuracy": 0.9901215243339538,
+      "eval_num_tokens": 13258191.0,
+      "eval_runtime": 20.2919,
+      "eval_samples_per_second": 36.961,
+      "eval_steps_per_second": 18.48,
+      "step": 2500
+    },
+    {
+      "entropy": 0.027981107782488836,
+      "epoch": 0.27015075376884423,
+      "grad_norm": 2.140625,
+      "learning_rate": 1.7127913325522546e-05,
+      "loss": 0.027564391493797302,
+      "mean_token_accuracy": 0.9884749103337527,
+      "num_tokens": 13360032.0,
+      "step": 2520
+    },
+    {
+      "entropy": 0.024831563800216828,
+      "epoch": 0.27229480737018424,
+      "grad_norm": 2.625,
+      "learning_rate": 1.7079041742888023e-05,
+      "loss": 0.027046608924865722,
+      "mean_token_accuracy": 0.9889734648168087,
+      "num_tokens": 13459997.0,
+      "step": 2540
+    },
+    {
+      "entropy": 0.025509285347288822,
+      "epoch": 0.2744388609715243,
+      "grad_norm": 2.875,
+      "learning_rate": 1.702982886418254e-05,
+      "loss": 0.026221618056297302,
+      "mean_token_accuracy": 0.9900430839508771,
+      "num_tokens": 13562047.0,
+      "step": 2560
+    },
+    {
+      "entropy": 0.02474073741786924,
+      "epoch": 0.2765829145728643,
+      "grad_norm": 2.875,
+      "learning_rate": 1.6980277062066513e-05,
+      "loss": 0.02481588125228882,
+      "mean_token_accuracy": 0.990446874499321,
+      "num_tokens": 13659557.0,
+      "step": 2580
+    },
+    {
+      "entropy": 0.026416731743847777,
+      "epoch": 0.27872696817420434,
+      "grad_norm": 1.7734375,
+      "learning_rate": 1.69303887255406e-05,
+      "loss": 0.026889228820800783,
+      "mean_token_accuracy": 0.9888020649552345,
+      "num_tokens": 13773310.0,
+      "step": 2600
+    },
+    {
+      "entropy": 0.0239254101291408,
+      "epoch": 0.2808710217755444,
+      "grad_norm": 1.4375,
+      "learning_rate": 1.6880166259830516e-05,
+      "loss": 0.02351238876581192,
+      "mean_token_accuracy": 0.9907949134707451,
+      "num_tokens": 13893984.0,
+      "step": 2620
+    },
+    {
+      "entropy": 0.02596387622606926,
+      "epoch": 0.2830150753768844,
+      "grad_norm": 1.859375,
+      "learning_rate": 1.682961208627108e-05,
+      "loss": 0.025493481755256654,
+      "mean_token_accuracy": 0.9891014736145735,
+      "num_tokens": 13994782.0,
+      "step": 2640
+    },
+    {
+      "entropy": 0.023969465121626853,
+      "epoch": 0.28515912897822443,
+      "grad_norm": 4.21875,
+      "learning_rate": 1.6778728642189456e-05,
+      "loss": 0.025132390856742858,
+      "mean_token_accuracy": 0.9904550548642874,
+      "num_tokens": 14106793.0,
+      "step": 2660
+    },
+    {
+      "entropy": 0.023048270364415657,
+      "epoch": 0.2873031825795645,
+      "grad_norm": 3.390625,
+      "learning_rate": 1.6727518380787678e-05,
+      "loss": 0.026464682817459107,
+      "mean_token_accuracy": 0.98928731828928,
+      "num_tokens": 14215090.0,
+      "step": 2680
+    },
+    {
+      "entropy": 0.024989640445710393,
+      "epoch": 0.2894472361809045,
+      "grad_norm": 1.0546875,
+      "learning_rate": 1.6675983771024345e-05,
+      "loss": 0.02458348125219345,
+      "mean_token_accuracy": 0.9901158206164837,
+      "num_tokens": 14321065.0,
+      "step": 2700
+    },
+    {
+      "entropy": 0.023273660653694607,
+      "epoch": 0.2915912897822446,
+      "grad_norm": 2.203125,
+      "learning_rate": 1.6624127297495606e-05,
+      "loss": 0.02475869804620743,
+      "mean_token_accuracy": 0.9907850239425897,
+      "num_tokens": 14421911.0,
+      "step": 2720
+    },
+    {
+      "entropy": 0.02422142572395387,
+      "epoch": 0.2937353433835846,
+      "grad_norm": 2.015625,
+      "learning_rate": 1.6571951460315362e-05,
+      "loss": 0.027780982851982116,
+      "mean_token_accuracy": 0.9886230506002903,
+      "num_tokens": 14529639.0,
+      "step": 2740
+    },
+    {
+      "entropy": 0.0267424136016416,
+      "epoch": 0.2958793969849246,
+      "grad_norm": 1.9765625,
+      "learning_rate": 1.6519458774994728e-05,
+      "loss": 0.026475533843040466,
+      "mean_token_accuracy": 0.9882622588425874,
+      "num_tokens": 14634262.0,
+      "step": 2760
+    },
+    {
+      "entropy": 0.023857305470937718,
+      "epoch": 0.29802345058626467,
+      "grad_norm": 2.390625,
+      "learning_rate": 1.646665177232077e-05,
+      "loss": 0.026801908016204835,
+      "mean_token_accuracy": 0.9902123864740133,
+      "num_tokens": 14739645.0,
+      "step": 2780
+    },
+    {
+      "entropy": 0.027312484692993168,
+      "epoch": 0.3001675041876047,
+      "grad_norm": 1.9765625,
+      "learning_rate": 1.6413532998234477e-05,
+      "loss": 0.027185559272766113,
+      "mean_token_accuracy": 0.9890467882156372,
+      "num_tokens": 14844155.0,
+      "step": 2800
+    },
+    {
+      "entropy": 0.02615051964248778,
+      "epoch": 0.3023115577889447,
+      "grad_norm": 1.859375,
+      "learning_rate": 1.6360105013708022e-05,
+      "loss": 0.02619759738445282,
+      "mean_token_accuracy": 0.9885728180408477,
+      "num_tokens": 14947978.0,
+      "step": 2820
+    },
+    {
+      "entropy": 0.023987949729507817,
+      "epoch": 0.30445561139028476,
+      "grad_norm": 1.2734375,
+      "learning_rate": 1.6306370394621274e-05,
+      "loss": 0.02713790535926819,
+      "mean_token_accuracy": 0.989062811806798,
+      "num_tokens": 15043851.0,
+      "step": 2840
+    },
+    {
+      "entropy": 0.024998777841574338,
+      "epoch": 0.3065996649916248,
+      "grad_norm": 2.59375,
+      "learning_rate": 1.625233173163765e-05,
+      "loss": 0.02687191665172577,
+      "mean_token_accuracy": 0.989005958288908,
+      "num_tokens": 15168916.0,
+      "step": 2860
+    },
+    {
+      "entropy": 0.025794855099957204,
+      "epoch": 0.30874371859296484,
+      "grad_norm": 1.296875,
+      "learning_rate": 1.619799163007916e-05,
+      "loss": 0.02666613459587097,
+      "mean_token_accuracy": 0.989616809785366,
+      "num_tokens": 15283048.0,
+      "step": 2880
+    },
+    {
+      "entropy": 0.023574712648041896,
+      "epoch": 0.31088777219430486,
+      "grad_norm": 1.2578125,
+      "learning_rate": 1.6143352709800847e-05,
+      "loss": 0.022058571875095367,
+      "mean_token_accuracy": 0.9904920358210803,
+      "num_tokens": 15401868.0,
+      "step": 2900
+    },
+    {
+      "entropy": 0.02349452629246116,
+      "epoch": 0.31303182579564487,
+      "grad_norm": 2.375,
+      "learning_rate": 1.6088417605064453e-05,
+      "loss": 0.025521630048751832,
+      "mean_token_accuracy": 0.9904454626142979,
+      "num_tokens": 15512377.0,
+      "step": 2920
+    },
+    {
+      "entropy": 0.024990445605726563,
+      "epoch": 0.31517587939698494,
+      "grad_norm": 1.7734375,
+      "learning_rate": 1.6033188964411407e-05,
+      "loss": 0.027597138285636903,
+      "mean_token_accuracy": 0.9910409476608038,
+      "num_tokens": 15626267.0,
+      "step": 2940
+    },
+    {
+      "entropy": 0.022828791862048094,
+      "epoch": 0.31731993299832495,
+      "grad_norm": 3.65625,
+      "learning_rate": 1.5977669450535164e-05,
+      "loss": 0.023614174127578734,
+      "mean_token_accuracy": 0.9908586539328098,
+      "num_tokens": 15732131.0,
+      "step": 2960
+    },
+    {
+      "entropy": 0.025171676824174937,
+      "epoch": 0.319463986599665,
+      "grad_norm": 1.0546875,
+      "learning_rate": 1.5921861740152803e-05,
+      "loss": 0.029384702444076538,
+      "mean_token_accuracy": 0.9898778181523085,
+      "num_tokens": 15827873.0,
+      "step": 2980
+    },
+    {
+      "entropy": 0.025553616087654517,
+      "epoch": 0.32160804020100503,
+      "grad_norm": 1.8828125,
+      "learning_rate": 1.586576852387599e-05,
+      "loss": 0.022602304816246033,
+      "mean_token_accuracy": 0.9912034373730421,
+      "num_tokens": 15920514.0,
+      "step": 3000
+    },
+    {
+      "epoch": 0.32160804020100503,
+      "eval_entropy": 0.024538942879318106,
+      "eval_loss": 0.02338256873190403,
+      "eval_mean_token_accuracy": 0.9907407927513122,
+      "eval_num_tokens": 15920514.0,
+      "eval_runtime": 20.2778,
+      "eval_samples_per_second": 36.986,
+      "eval_steps_per_second": 18.493,
+      "step": 3000
+    },
+    {
+      "entropy": 0.021682905555485375,
+      "epoch": 0.32375209380234504,
+      "grad_norm": 2.46875,
+      "learning_rate": 1.5809392506081244e-05,
+      "loss": 0.022935517132282257,
+      "mean_token_accuracy": 0.9904364123940468,
+      "num_tokens": 16035334.0,
+      "step": 3020
+    },
+    {
+      "entropy": 0.01898658178452024,
+      "epoch": 0.3258961474036851,
+      "grad_norm": 1.6796875,
+      "learning_rate": 1.5752736404779577e-05,
+      "loss": 0.02092217206954956,
+      "mean_token_accuracy": 0.984833637624979,
+      "num_tokens": 16140327.0,
+      "step": 3040
+    },
+    {
+      "entropy": 0.021216248062864908,
+      "epoch": 0.3280402010050251,
+      "grad_norm": 2.3125,
+      "learning_rate": 1.5695802951485434e-05,
+      "loss": 0.024958278238773345,
+      "mean_token_accuracy": 0.9900662276893855,
+      "num_tokens": 16255099.0,
+      "step": 3060
+    },
+    {
+      "entropy": 0.023438269265625422,
+      "epoch": 0.33018425460636514,
+      "grad_norm": 1.7421875,
+      "learning_rate": 1.5638594891085e-05,
+      "loss": 0.024300065636634827,
+      "mean_token_accuracy": 0.9898340333253145,
+      "num_tokens": 16366533.0,
+      "step": 3080
+    },
+    {
+      "entropy": 0.02112419887921533,
+      "epoch": 0.3323283082077052,
+      "grad_norm": 1.5078125,
+      "learning_rate": 1.5581114981703878e-05,
+      "loss": 0.02245298624038696,
+      "mean_token_accuracy": 0.9914943426847458,
+      "num_tokens": 16469428.0,
+      "step": 3100
+    },
+    {
+      "entropy": 0.023446582856286112,
+      "epoch": 0.3344723618090452,
+      "grad_norm": 2.78125,
+      "learning_rate": 1.5523365994574098e-05,
+      "loss": 0.024820271134376525,
+      "mean_token_accuracy": 0.990750040858984,
+      "num_tokens": 16582498.0,
+      "step": 3120
+    },
+    {
+      "entropy": 0.022754177751903627,
+      "epoch": 0.3366164154103853,
+      "grad_norm": 1.8203125,
+      "learning_rate": 1.546535071390052e-05,
+      "loss": 0.025619438290596007,
+      "mean_token_accuracy": 0.9895062685012818,
+      "num_tokens": 16693045.0,
+      "step": 3140
+    },
+    {
+      "entropy": 0.023310942134048673,
+      "epoch": 0.3387604690117253,
+      "grad_norm": 2.546875,
+      "learning_rate": 1.540707193672659e-05,
+      "loss": 0.02326971888542175,
+      "mean_token_accuracy": 0.9908227741718292,
+      "num_tokens": 16796636.0,
+      "step": 3160
+    },
+    {
+      "entropy": 0.02383078648617811,
+      "epoch": 0.3409045226130653,
+      "grad_norm": 3.390625,
+      "learning_rate": 1.53485324727995e-05,
+      "loss": 0.025269970297813416,
+      "mean_token_accuracy": 0.9894092302769423,
+      "num_tokens": 16908823.0,
+      "step": 3180
+    },
+    {
+      "entropy": 0.025040690238802198,
+      "epoch": 0.3430485762144054,
+      "grad_norm": 2.515625,
+      "learning_rate": 1.5289735144434733e-05,
+      "loss": 0.025703960657119752,
+      "mean_token_accuracy": 0.9828102149069309,
+      "num_tokens": 17011175.0,
+      "step": 3200
+    },
+    {
+      "entropy": 0.0229463691044657,
+      "epoch": 0.3451926298157454,
+      "grad_norm": 3.9375,
+      "learning_rate": 1.5230682786379958e-05,
+      "loss": 0.0237918958067894,
+      "mean_token_accuracy": 0.9901813805103302,
+      "num_tokens": 17118025.0,
+      "step": 3220
+    },
+    {
+      "entropy": 0.02446629966070759,
+      "epoch": 0.3473366834170854,
+      "grad_norm": 1.3359375,
+      "learning_rate": 1.5171378245678398e-05,
+      "loss": 0.025769346952438356,
+      "mean_token_accuracy": 0.988752768561244,
+      "num_tokens": 17225787.0,
+      "step": 3240
+    },
+    {
+      "entropy": 0.023929561218301386,
+      "epoch": 0.34948073701842547,
+      "grad_norm": 2.890625,
+      "learning_rate": 1.5111824381531543e-05,
+      "loss": 0.02350829541683197,
+      "mean_token_accuracy": 0.9899691499769687,
+      "num_tokens": 17326334.0,
+      "step": 3260
+    },
+    {
+      "entropy": 0.023094384433488813,
+      "epoch": 0.3516247906197655,
+      "grad_norm": 1.3125,
+      "learning_rate": 1.5052024065161321e-05,
+      "loss": 0.025211021304130554,
+      "mean_token_accuracy": 0.989602592214942,
+      "num_tokens": 17427911.0,
+      "step": 3280
+    },
+    {
+      "entropy": 0.02580152012506005,
+      "epoch": 0.35376884422110555,
+      "grad_norm": 0.88671875,
+      "learning_rate": 1.4991980179671656e-05,
+      "loss": 0.028021180629730226,
+      "mean_token_accuracy": 0.9875154685229063,
+      "num_tokens": 17535492.0,
+      "step": 3300
+    },
+    {
+      "entropy": 0.025954745234002985,
+      "epoch": 0.35591289782244556,
+      "grad_norm": 1.46875,
+      "learning_rate": 1.4931695619909466e-05,
+      "loss": 0.023851674795150758,
+      "mean_token_accuracy": 0.9898436293005943,
+      "num_tokens": 17641670.0,
+      "step": 3320
+    },
+    {
+      "entropy": 0.024273743057074172,
+      "epoch": 0.3580569514237856,
+      "grad_norm": 1.7109375,
+      "learning_rate": 1.4871173292325106e-05,
+      "loss": 0.02511066496372223,
+      "mean_token_accuracy": 0.989355530589819,
+      "num_tokens": 17742555.0,
+      "step": 3340
+    },
+    {
+      "entropy": 0.02353285207400404,
+      "epoch": 0.36020100502512564,
+      "grad_norm": 1.4921875,
+      "learning_rate": 1.4810416114832236e-05,
+      "loss": 0.02515583038330078,
+      "mean_token_accuracy": 0.9906750213354826,
+      "num_tokens": 17848901.0,
+      "step": 3360
+    },
+    {
+      "entropy": 0.022243685176727014,
+      "epoch": 0.36234505862646565,
+      "grad_norm": 2.453125,
+      "learning_rate": 1.474942701666715e-05,
+      "loss": 0.02309694290161133,
+      "mean_token_accuracy": 0.9912830431014299,
+      "num_tokens": 17955457.0,
+      "step": 3380
+    },
+    {
+      "entropy": 0.022329559885474737,
+      "epoch": 0.3644891122278057,
+      "grad_norm": 1.6171875,
+      "learning_rate": 1.4688208938247535e-05,
+      "loss": 0.024044957756996155,
+      "mean_token_accuracy": 0.9913502413779497,
+      "num_tokens": 18054439.0,
+      "step": 3400
+    },
+    {
+      "entropy": 0.023860087489720173,
+      "epoch": 0.36663316582914574,
+      "grad_norm": 1.84375,
+      "learning_rate": 1.462676483103072e-05,
+      "loss": 0.023865839838981627,
+      "mean_token_accuracy": 0.9902952261269092,
+      "num_tokens": 18159108.0,
+      "step": 3420
+    },
+    {
+      "entropy": 0.021247414961362666,
+      "epoch": 0.36877721943048575,
+      "grad_norm": 1.8359375,
+      "learning_rate": 1.4565097657371378e-05,
+      "loss": 0.021402713656425477,
+      "mean_token_accuracy": 0.9921348176896572,
+      "num_tokens": 18268330.0,
+      "step": 3440
+    },
+    {
+      "entropy": 0.021685771951888456,
+      "epoch": 0.3709212730318258,
+      "grad_norm": 1.2109375,
+      "learning_rate": 1.4503210390378701e-05,
+      "loss": 0.026408180594444275,
+      "mean_token_accuracy": 0.9894901543855668,
+      "num_tokens": 18376382.0,
+      "step": 3460
+    },
+    {
+      "entropy": 0.025453939728959086,
+      "epoch": 0.37306532663316583,
+      "grad_norm": 2.09375,
+      "learning_rate": 1.444110601377306e-05,
+      "loss": 0.026739346981048583,
+      "mean_token_accuracy": 0.9884634926915169,
+      "num_tokens": 18489633.0,
+      "step": 3480
+    },
+    {
+      "entropy": 0.02501170370624095,
+      "epoch": 0.37520938023450584,
+      "grad_norm": 2.03125,
+      "learning_rate": 1.437878752174216e-05,
+      "loss": 0.02536824345588684,
+      "mean_token_accuracy": 0.9898253668099641,
+      "num_tokens": 18597861.0,
+      "step": 3500
+    },
+    {
+      "epoch": 0.37520938023450584,
+      "eval_entropy": 0.0244218321183871,
+      "eval_loss": 0.02227754145860672,
+      "eval_mean_token_accuracy": 0.9911262426376343,
+      "eval_num_tokens": 18597861.0,
+      "eval_runtime": 20.324,
+      "eval_samples_per_second": 36.902,
+      "eval_steps_per_second": 18.451,
+      "step": 3500
+    },
+    {
+      "entropy": 0.0246668197542931,
+      "epoch": 0.3773534338358459,
+      "grad_norm": 1.7421875,
+      "learning_rate": 1.4316257918796671e-05,
+      "loss": 0.025204175710678102,
+      "mean_token_accuracy": 0.99043660312891,
+      "num_tokens": 18697672.0,
+      "step": 3520
+    },
+    {
+      "entropy": 0.022950895108533587,
+      "epoch": 0.3794974874371859,
+      "grad_norm": 1.828125,
+      "learning_rate": 1.4253520219625387e-05,
+      "loss": 0.02428869903087616,
+      "mean_token_accuracy": 0.9907822337001562,
+      "num_tokens": 18811789.0,
+      "step": 3540
+    },
+    {
+      "entropy": 0.023854929314802577,
+      "epoch": 0.381641541038526,
+      "grad_norm": 1.9609375,
+      "learning_rate": 1.4190577448949876e-05,
+      "loss": 0.023869043588638304,
+      "mean_token_accuracy": 0.9913762111216784,
+      "num_tokens": 18911590.0,
+      "step": 3560
+    },
+    {
+      "entropy": 0.024182767125876126,
+      "epoch": 0.383785594639866,
+      "grad_norm": 2.546875,
+      "learning_rate": 1.4127432641378649e-05,
+      "loss": 0.023478889465332033,
+      "mean_token_accuracy": 0.9907276172190904,
+      "num_tokens": 19026734.0,
+      "step": 3580
+    },
+    {
+      "entropy": 0.020359245626377743,
+      "epoch": 0.385929648241206,
+      "grad_norm": 1.6640625,
+      "learning_rate": 1.4064088841260857e-05,
+      "loss": 0.023241102695465088,
+      "mean_token_accuracy": 0.9907599698752165,
+      "num_tokens": 19129417.0,
+      "step": 3600
+    },
+    {
+      "entropy": 0.021383443258810076,
+      "epoch": 0.3880737018425461,
+      "grad_norm": 1.3203125,
+      "learning_rate": 1.4000549102539516e-05,
+      "loss": 0.023419806361198427,
+      "mean_token_accuracy": 0.990417006239295,
+      "num_tokens": 19235205.0,
+      "step": 3620
+    },
+    {
+      "entropy": 0.020958786468827383,
+      "epoch": 0.3902177554438861,
+      "grad_norm": 1.046875,
+      "learning_rate": 1.3936816488604273e-05,
+      "loss": 0.02310475558042526,
+      "mean_token_accuracy": 0.991861392185092,
+      "num_tokens": 19333977.0,
+      "step": 3640
+    },
+    {
+      "entropy": 0.022316582887765436,
+      "epoch": 0.3923618090452261,
+      "grad_norm": 2.84375,
+      "learning_rate": 1.3872894072143708e-05,
+      "loss": 0.02316024452447891,
+      "mean_token_accuracy": 0.9916662760078907,
+      "num_tokens": 19433681.0,
+      "step": 3660
+    },
+    {
+      "entropy": 0.023413096193553428,
+      "epoch": 0.3945058626465662,
+      "grad_norm": 1.765625,
+      "learning_rate": 1.3808784934997199e-05,
+      "loss": 0.026311498880386353,
+      "mean_token_accuracy": 0.9898762315511703,
+      "num_tokens": 19543600.0,
+      "step": 3680
+    },
+    {
+      "entropy": 0.02430928884550667,
+      "epoch": 0.3966499162479062,
+      "grad_norm": 1.640625,
+      "learning_rate": 1.3744492168006332e-05,
+      "loss": 0.023604103922843934,
+      "mean_token_accuracy": 0.9905571412295103,
+      "num_tokens": 19656000.0,
+      "step": 3700
+    },
+    {
+      "entropy": 0.02260660900547009,
+      "epoch": 0.39879396984924625,
+      "grad_norm": 2.078125,
+      "learning_rate": 1.368001887086589e-05,
+      "loss": 0.024041768908500672,
+      "mean_token_accuracy": 0.9895115055143833,
+      "num_tokens": 19756250.0,
+      "step": 3720
+    },
+    {
+      "entropy": 0.02264130887056126,
+      "epoch": 0.40093802345058627,
+      "grad_norm": 1.625,
+      "learning_rate": 1.361536815197441e-05,
+      "loss": 0.025104725360870363,
+      "mean_token_accuracy": 0.9906443733721971,
+      "num_tokens": 19878359.0,
+      "step": 3740
+    },
+    {
+      "entropy": 0.023430266666673562,
+      "epoch": 0.4030820770519263,
+      "grad_norm": 1.65625,
+      "learning_rate": 1.3550543128284323e-05,
+      "loss": 0.024283410608768465,
+      "mean_token_accuracy": 0.9907973647117615,
+      "num_tokens": 19972202.0,
+      "step": 3760
+    },
+    {
+      "entropy": 0.02248989698346122,
+      "epoch": 0.40522613065326635,
+      "grad_norm": 1.4609375,
+      "learning_rate": 1.348554692515167e-05,
+      "loss": 0.02604263722896576,
+      "mean_token_accuracy": 0.9889465566724539,
+      "num_tokens": 20076482.0,
+      "step": 3780
+    },
+    {
+      "entropy": 0.02452128016102506,
+      "epoch": 0.40737018425460636,
+      "grad_norm": 1.703125,
+      "learning_rate": 1.3420382676185421e-05,
+      "loss": 0.025164645910263062,
+      "mean_token_accuracy": 0.9902698881924152,
+      "num_tokens": 20186450.0,
+      "step": 3800
+    },
+    {
+      "entropy": 0.024925458200004868,
+      "epoch": 0.4095142378559464,
+      "grad_norm": 1.921875,
+      "learning_rate": 1.335505352309642e-05,
+      "loss": 0.025616466999053955,
+      "mean_token_accuracy": 0.989701810106635,
+      "num_tokens": 20290995.0,
+      "step": 3820
+    },
+    {
+      "entropy": 0.022864083916329037,
+      "epoch": 0.41165829145728644,
+      "grad_norm": 1.8125,
+      "learning_rate": 1.3289562615545883e-05,
+      "loss": 0.02312360405921936,
+      "mean_token_accuracy": 0.9906341150403023,
+      "num_tokens": 20392839.0,
+      "step": 3840
+    },
+    {
+      "entropy": 0.022927879701637722,
+      "epoch": 0.41380234505862645,
+      "grad_norm": 1.2421875,
+      "learning_rate": 1.3223913110993576e-05,
+      "loss": 0.02285770922899246,
+      "mean_token_accuracy": 0.9910362754017115,
+      "num_tokens": 20499794.0,
+      "step": 3860
+    },
+    {
+      "entropy": 0.02117061285384807,
+      "epoch": 0.4159463986599665,
+      "grad_norm": 11.0,
+      "learning_rate": 1.3158108174545556e-05,
+      "loss": 0.021062937378883363,
+      "mean_token_accuracy": 0.9917857561260461,
+      "num_tokens": 20612104.0,
+      "step": 3880
+    },
+    {
+      "entropy": 0.02190584047739321,
+      "epoch": 0.41809045226130653,
+      "grad_norm": 2.625,
+      "learning_rate": 1.3092150978801605e-05,
+      "loss": 0.023377425968647003,
+      "mean_token_accuracy": 0.9904357828199863,
+      "num_tokens": 20723068.0,
+      "step": 3900
+    },
+    {
+      "entropy": 0.02153953154429473,
+      "epoch": 0.42023450586264655,
+      "grad_norm": 2.703125,
+      "learning_rate": 1.3026044703702249e-05,
+      "loss": 0.023467777669429778,
+      "mean_token_accuracy": 0.9905689381062984,
+      "num_tokens": 20826557.0,
+      "step": 3920
+    },
+    {
+      "entropy": 0.02152182651061594,
+      "epoch": 0.4223785594639866,
+      "grad_norm": 1.671875,
+      "learning_rate": 1.2959792536375471e-05,
+      "loss": 0.020291084051132204,
+      "mean_token_accuracy": 0.992015489563346,
+      "num_tokens": 20929814.0,
+      "step": 3940
+    },
+    {
+      "entropy": 0.02025915689789599,
+      "epoch": 0.4245226130653266,
+      "grad_norm": 1.7421875,
+      "learning_rate": 1.2893397670983022e-05,
+      "loss": 0.020092053711414336,
+      "mean_token_accuracy": 0.9926223222166299,
+      "num_tokens": 21032536.0,
+      "step": 3960
+    },
+    {
+      "entropy": 0.019897875330707394,
+      "epoch": 0.4266666666666667,
+      "grad_norm": 2.46875,
+      "learning_rate": 1.2826863308566437e-05,
+      "loss": 0.022923128306865694,
+      "mean_token_accuracy": 0.9915768329054118,
+      "num_tokens": 21125010.0,
+      "step": 3980
+    },
+    {
+      "entropy": 0.020706827863295984,
+      "epoch": 0.4288107202680067,
+      "grad_norm": 1.9140625,
+      "learning_rate": 1.2760192656892723e-05,
+      "loss": 0.022419747710227967,
+      "mean_token_accuracy": 0.9913075774908066,
+      "num_tokens": 21242867.0,
+      "step": 4000
+    },
+    {
+      "epoch": 0.4288107202680067,
+      "eval_entropy": 0.01964530904859809,
+      "eval_loss": 0.02178012579679489,
+      "eval_mean_token_accuracy": 0.9911314179102579,
+      "eval_num_tokens": 21242867.0,
+      "eval_runtime": 20.3146,
+      "eval_samples_per_second": 36.919,
+      "eval_steps_per_second": 18.46,
+      "step": 4000
+    },
+    {
+      "entropy": 0.02013387772437909,
+      "epoch": 0.4309547738693467,
+      "grad_norm": 1.5078125,
+      "learning_rate": 1.2693388930299672e-05,
+      "loss": 0.021522065997123717,
+      "mean_token_accuracy": 0.9929842289537192,
+      "num_tokens": 21346530.0,
+      "step": 4020
+    },
+    {
+      "entropy": 0.0202242263831522,
+      "epoch": 0.4330988274706868,
+      "grad_norm": 2.828125,
+      "learning_rate": 1.2626455349540918e-05,
+      "loss": 0.022661156952381134,
+      "mean_token_accuracy": 0.9919143192470073,
+      "num_tokens": 21458250.0,
+      "step": 4040
+    },
+    {
+      "entropy": 0.022103192698341444,
+      "epoch": 0.4352428810720268,
+      "grad_norm": 1.3671875,
+      "learning_rate": 1.2559395141630656e-05,
+      "loss": 0.02430972158908844,
+      "mean_token_accuracy": 0.9904460556805134,
+      "num_tokens": 21563008.0,
+      "step": 4060
+    },
+    {
+      "entropy": 0.02260455262266987,
+      "epoch": 0.4373869346733668,
+      "grad_norm": 1.4609375,
+      "learning_rate": 1.2492211539688036e-05,
+      "loss": 0.025785067677497865,
+      "mean_token_accuracy": 0.9895172834396362,
+      "num_tokens": 21668155.0,
+      "step": 4080
+    },
+    {
+      "entropy": 0.02297381363787281,
+      "epoch": 0.4395309882747069,
+      "grad_norm": 2.203125,
+      "learning_rate": 1.2424907782781313e-05,
+      "loss": 0.024609476327896118,
+      "mean_token_accuracy": 0.990400804951787,
+      "num_tokens": 21784031.0,
+      "step": 4100
+    },
+    {
+      "entropy": 0.022381592743295188,
+      "epoch": 0.4416750418760469,
+      "grad_norm": 1.609375,
+      "learning_rate": 1.2357487115771673e-05,
+      "loss": 0.023763097822666168,
+      "mean_token_accuracy": 0.99109130166471,
+      "num_tokens": 21893847.0,
+      "step": 4120
+    },
+    {
+      "entropy": 0.022086531718105106,
+      "epoch": 0.44381909547738696,
+      "grad_norm": 2.0,
+      "learning_rate": 1.2289952789156798e-05,
+      "loss": 0.020330893993377685,
+      "mean_token_accuracy": 0.9921020328998565,
+      "num_tokens": 22005454.0,
+      "step": 4140
+    },
+    {
+      "entropy": 0.02248057241056358,
+      "epoch": 0.445963149078727,
+      "grad_norm": 1.609375,
+      "learning_rate": 1.222230805891414e-05,
+      "loss": 0.02294687330722809,
+      "mean_token_accuracy": 0.9910975113511086,
+      "num_tokens": 22107232.0,
+      "step": 4160
+    },
+    {
+      "entropy": 0.02265706049165601,
+      "epoch": 0.448107202680067,
+      "grad_norm": 3.90625,
+      "learning_rate": 1.215455618634396e-05,
+      "loss": 0.025840884447097777,
+      "mean_token_accuracy": 0.9897211868315935,
+      "num_tokens": 22217548.0,
+      "step": 4180
+    },
+    {
+      "entropy": 0.02281195720179312,
+      "epoch": 0.45025125628140705,
+      "grad_norm": 3.1875,
+      "learning_rate": 1.2086700437912077e-05,
+      "loss": 0.024470710754394533,
+      "mean_token_accuracy": 0.9903297979384661,
+      "num_tokens": 22328036.0,
+      "step": 4200
+    },
+    {
+      "entropy": 0.02147264582108619,
+      "epoch": 0.45239530988274707,
+      "grad_norm": 1.6328125,
+      "learning_rate": 1.2018744085092399e-05,
+      "loss": 0.021890904009342193,
+      "mean_token_accuracy": 0.9921745974570513,
+      "num_tokens": 22439360.0,
+      "step": 4220
+    },
+    {
+      "entropy": 0.021090176403095028,
+      "epoch": 0.4545393634840871,
+      "grad_norm": 1.6171875,
+      "learning_rate": 1.1950690404209186e-05,
+      "loss": 0.022462236881256103,
+      "mean_token_accuracy": 0.992034525051713,
+      "num_tokens": 22543357.0,
+      "step": 4240
+    },
+    {
+      "entropy": 0.022680416493108167,
+      "epoch": 0.45668341708542715,
+      "grad_norm": 1.328125,
+      "learning_rate": 1.1882542676279098e-05,
+      "loss": 0.02402680218219757,
+      "mean_token_accuracy": 0.9908802334219218,
+      "num_tokens": 22650751.0,
+      "step": 4260
+    },
+    {
+      "entropy": 0.0232969627111288,
+      "epoch": 0.45882747068676716,
+      "grad_norm": 2.96875,
+      "learning_rate": 1.1814304186853009e-05,
+      "loss": 0.022290503978729247,
+      "mean_token_accuracy": 0.991101274266839,
+      "num_tokens": 22753548.0,
+      "step": 4280
+    },
+    {
+      "entropy": 0.021912004817795604,
+      "epoch": 0.4609715242881072,
+      "grad_norm": 3.359375,
+      "learning_rate": 1.1745978225857609e-05,
+      "loss": 0.022448597848415373,
+      "mean_token_accuracy": 0.9917845461517573,
+      "num_tokens": 22861310.0,
+      "step": 4300
+    },
+    {
+      "entropy": 0.021719314875508645,
+      "epoch": 0.46311557788944724,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.1677568087436767e-05,
+      "loss": 0.023140157759189605,
+      "mean_token_accuracy": 0.9908079355955124,
+      "num_tokens": 22978927.0,
+      "step": 4320
+    },
+    {
+      "entropy": 0.021950446777191247,
+      "epoch": 0.46525963149078725,
+      "grad_norm": 2.203125,
+      "learning_rate": 1.1609077069792754e-05,
+      "loss": 0.0213756799697876,
+      "mean_token_accuracy": 0.9913494195789099,
+      "num_tokens": 23082226.0,
+      "step": 4340
+    },
+    {
+      "entropy": 0.02140526709590631,
+      "epoch": 0.4674036850921273,
+      "grad_norm": 2.15625,
+      "learning_rate": 1.1540508475027179e-05,
+      "loss": 0.020155489444732666,
+      "mean_token_accuracy": 0.9919397257268429,
+      "num_tokens": 23180655.0,
+      "step": 4360
+    },
+    {
+      "entropy": 0.017593615273926843,
+      "epoch": 0.46954773869346733,
+      "grad_norm": 1.75,
+      "learning_rate": 1.147186560898183e-05,
+      "loss": 0.02236589938402176,
+      "mean_token_accuracy": 0.9917930167168378,
+      "num_tokens": 23283721.0,
+      "step": 4380
+    },
+    {
+      "entropy": 0.020877945362190075,
+      "epoch": 0.47169179229480734,
+      "grad_norm": 1.625,
+      "learning_rate": 1.1403151781079278e-05,
+      "loss": 0.022862054407596588,
+      "mean_token_accuracy": 0.9906388968229294,
+      "num_tokens": 23390288.0,
+      "step": 4400
+    },
+    {
+      "entropy": 0.01995794570948419,
+      "epoch": 0.4738358458961474,
+      "grad_norm": 2.109375,
+      "learning_rate": 1.1334370304163306e-05,
+      "loss": 0.02168610394001007,
+      "mean_token_accuracy": 0.9925977524369956,
+      "num_tokens": 23485464.0,
+      "step": 4420
+    },
+    {
+      "entropy": 0.01932002087028195,
+      "epoch": 0.4759798994974874,
+      "grad_norm": 1.6796875,
+      "learning_rate": 1.1265524494339219e-05,
+      "loss": 0.02263712286949158,
+      "mean_token_accuracy": 0.9910555865615607,
+      "num_tokens": 23585708.0,
+      "step": 4440
+    },
+    {
+      "entropy": 0.021627131735112926,
+      "epoch": 0.4781239530988275,
+      "grad_norm": 2.140625,
+      "learning_rate": 1.1196617670813936e-05,
+      "loss": 0.022156818211078642,
+      "mean_token_accuracy": 0.9912719707936049,
+      "num_tokens": 23693455.0,
+      "step": 4460
+    },
+    {
+      "entropy": 0.02139281702184235,
+      "epoch": 0.4802680067001675,
+      "grad_norm": 2.1875,
+      "learning_rate": 1.1127653155735993e-05,
+      "loss": 0.022481553256511688,
+      "mean_token_accuracy": 0.9902749702334404,
+      "num_tokens": 23800058.0,
+      "step": 4480
+    },
+    {
+      "entropy": 0.02091277845497643,
+      "epoch": 0.4824120603015075,
+      "grad_norm": 2.328125,
+      "learning_rate": 1.1058634274035361e-05,
+      "loss": 0.02205648124217987,
+      "mean_token_accuracy": 0.9916612818837166,
+      "num_tokens": 23910439.0,
+      "step": 4500
+    },
+    {
+      "epoch": 0.4824120603015075,
+      "eval_entropy": 0.021846828331123107,
+      "eval_loss": 0.02092284895479679,
+      "eval_mean_token_accuracy": 0.9913039420445761,
+      "eval_num_tokens": 23910439.0,
+      "eval_runtime": 20.2669,
+      "eval_samples_per_second": 37.006,
+      "eval_steps_per_second": 18.503,
+      "step": 4500
+    },
+    {
+      "entropy": 0.021989129094345118,
+      "epoch": 0.4845561139028476,
+      "grad_norm": 1.765625,
+      "learning_rate": 1.0989564353263133e-05,
+      "loss": 0.02231614738702774,
+      "mean_token_accuracy": 0.9905249148607254,
+      "num_tokens": 24019503.0,
+      "step": 4520
+    },
+    {
+      "entropy": 0.02183805209115235,
+      "epoch": 0.4867001675041876,
+      "grad_norm": 2.3125,
+      "learning_rate": 1.0920446723431127e-05,
+      "loss": 0.02185622453689575,
+      "mean_token_accuracy": 0.9918582946062088,
+      "num_tokens": 24124480.0,
+      "step": 4540
+    },
+    {
+      "entropy": 0.02131056402449758,
+      "epoch": 0.48884422110552767,
+      "grad_norm": 1.3203125,
+      "learning_rate": 1.0851284716851295e-05,
+      "loss": 0.022905312478542328,
+      "mean_token_accuracy": 0.9910277899354696,
+      "num_tokens": 24230705.0,
+      "step": 4560
+    },
+    {
+      "entropy": 0.023312085462430333,
+      "epoch": 0.4909882747068677,
+      "grad_norm": 1.8828125,
+      "learning_rate": 1.0782081667975106e-05,
+      "loss": 0.022799497842788695,
+      "mean_token_accuracy": 0.9904710713773965,
+      "num_tokens": 24335812.0,
+      "step": 4580
+    },
+    {
+      "entropy": 0.02137677239184086,
+      "epoch": 0.4931323283082077,
+      "grad_norm": 1.34375,
+      "learning_rate": 1.0712840913232768e-05,
+      "loss": 0.02341741919517517,
+      "mean_token_accuracy": 0.9909932963550091,
+      "num_tokens": 24444839.0,
+      "step": 4600
+    },
+    {
+      "entropy": 0.022716682115378715,
+      "epoch": 0.49527638190954776,
+      "grad_norm": 1.484375,
+      "learning_rate": 1.0643565790872359e-05,
+      "loss": 0.02410309910774231,
+      "mean_token_accuracy": 0.9909039065241814,
+      "num_tokens": 24545937.0,
+      "step": 4620
+    },
+    {
+      "entropy": 0.022203819393803316,
+      "epoch": 0.49742043551088777,
+      "grad_norm": 1.8515625,
+      "learning_rate": 1.0574259640798908e-05,
+      "loss": 0.022010770440101624,
+      "mean_token_accuracy": 0.9920209400355816,
+      "num_tokens": 24647974.0,
+      "step": 4640
+    },
+    {
+      "entropy": 0.02109122319297967,
+      "epoch": 0.4995644891122278,
+      "grad_norm": 1.3359375,
+      "learning_rate": 1.0504925804413348e-05,
+      "loss": 0.02280532568693161,
+      "mean_token_accuracy": 0.9912972040474415,
+      "num_tokens": 24752530.0,
+      "step": 4660
+    },
+    {
+      "entropy": 0.022384429502744753,
+      "epoch": 0.5017085427135678,
+      "grad_norm": 2.390625,
+      "learning_rate": 1.0435567624451435e-05,
+      "loss": 0.027354520559310914,
+      "mean_token_accuracy": 0.98953907340765,
+      "num_tokens": 24854708.0,
+      "step": 4680
+    },
+    {
+      "entropy": 0.02178340100826972,
+      "epoch": 0.5038525963149079,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.0366188444822577e-05,
+      "loss": 0.02108224630355835,
+      "mean_token_accuracy": 0.991971093788743,
+      "num_tokens": 24951913.0,
+      "step": 4700
+    },
+    {
+      "entropy": 0.0228582585350523,
+      "epoch": 0.5059966499162479,
+      "grad_norm": 1.578125,
+      "learning_rate": 1.0296791610448634e-05,
+      "loss": 0.02327609956264496,
+      "mean_token_accuracy": 0.9911777719855308,
+      "num_tokens": 25057377.0,
+      "step": 4720
+    },
+    {
+      "entropy": 0.022045019802681053,
+      "epoch": 0.5081407035175879,
+      "grad_norm": 1.1953125,
+      "learning_rate": 1.0227380467102627e-05,
+      "loss": 0.023243658244609833,
+      "mean_token_accuracy": 0.9909518156200647,
+      "num_tokens": 25170778.0,
+      "step": 4740
+    },
+    {
+      "entropy": 0.01989219032784604,
+      "epoch": 0.510284757118928,
+      "grad_norm": 5.03125,
+      "learning_rate": 1.0157958361247448e-05,
+      "loss": 0.020330929756164552,
+      "mean_token_accuracy": 0.9921452704817056,
+      "num_tokens": 25284310.0,
+      "step": 4760
+    },
+    {
+      "entropy": 0.021722077413278383,
+      "epoch": 0.512428810720268,
+      "grad_norm": 1.6015625,
+      "learning_rate": 1.0088528639874516e-05,
+      "loss": 0.02167322486639023,
+      "mean_token_accuracy": 0.9913335714489222,
+      "num_tokens": 25392904.0,
+      "step": 4780
+    },
+    {
+      "entropy": 0.022090970205681515,
+      "epoch": 0.5145728643216081,
+      "grad_norm": 1.390625,
+      "learning_rate": 1.0019094650342409e-05,
+      "loss": 0.023602089285850524,
+      "mean_token_accuracy": 0.9912930846214294,
+      "num_tokens": 25505548.0,
+      "step": 4800
+    },
+    {
+      "entropy": 0.021318517662666635,
+      "epoch": 0.516716917922948,
+      "grad_norm": 1.8046875,
+      "learning_rate": 9.94965974021548e-06,
+      "loss": 0.021860493719577788,
+      "mean_token_accuracy": 0.9906282421201468,
+      "num_tokens": 25616943.0,
+      "step": 4820
+    },
+    {
+      "entropy": 0.020716092844122614,
+      "epoch": 0.5188609715242881,
+      "grad_norm": 1.7109375,
+      "learning_rate": 9.880227257102483e-06,
+      "loss": 0.021078214049339294,
+      "mean_token_accuracy": 0.9921130206435919,
+      "num_tokens": 25727510.0,
+      "step": 4840
+    },
+    {
+      "entropy": 0.02081386552129061,
+      "epoch": 0.5210050251256282,
+      "grad_norm": 2.203125,
+      "learning_rate": 9.810800548495132e-06,
+      "loss": 0.019788502156734465,
+      "mean_token_accuracy": 0.9920309089124203,
+      "num_tokens": 25831138.0,
+      "step": 4860
+    },
+    {
+      "entropy": 0.020538337337939083,
+      "epoch": 0.5231490787269681,
+      "grad_norm": 3.671875,
+      "learning_rate": 9.741382961606768e-06,
+      "loss": 0.021504290401935577,
+      "mean_token_accuracy": 0.9916285052895546,
+      "num_tokens": 25938256.0,
+      "step": 4880
+    },
+    {
+      "entropy": 0.02026532054192103,
+      "epoch": 0.5252931323283082,
+      "grad_norm": 2.015625,
+      "learning_rate": 9.671977843210928e-06,
+      "loss": 0.022037246823310853,
+      "mean_token_accuracy": 0.9909355752170086,
+      "num_tokens": 26045653.0,
+      "step": 4900
+    },
+    {
+      "entropy": 0.021552794340914262,
+      "epoch": 0.5274371859296483,
+      "grad_norm": 1.9453125,
+      "learning_rate": 9.602588539480037e-06,
+      "loss": 0.024588188529014586,
+      "mean_token_accuracy": 0.9902191430330276,
+      "num_tokens": 26147507.0,
+      "step": 4920
+    },
+    {
+      "entropy": 0.02167862878472988,
+      "epoch": 0.5295812395309882,
+      "grad_norm": 1.9921875,
+      "learning_rate": 9.533218395824046e-06,
+      "loss": 0.023571619391441347,
+      "mean_token_accuracy": 0.9900527201592922,
+      "num_tokens": 26257125.0,
+      "step": 4940
+    },
+    {
+      "entropy": 0.02021576279623787,
+      "epoch": 0.5317252931323283,
+      "grad_norm": 2.1875,
+      "learning_rate": 9.463870756729163e-06,
+      "loss": 0.019122126698493957,
+      "mean_token_accuracy": 0.992604573443532,
+      "num_tokens": 26364694.0,
+      "step": 4960
+    },
+    {
+      "entropy": 0.019900748710369953,
+      "epoch": 0.5338693467336684,
+      "grad_norm": 2.0625,
+      "learning_rate": 9.394548965596609e-06,
+      "loss": 0.01807314157485962,
+      "mean_token_accuracy": 0.9929017867892981,
+      "num_tokens": 26473983.0,
+      "step": 4980
+    },
+    {
+      "entropy": 0.0207442283473938,
+      "epoch": 0.5360134003350083,
+      "grad_norm": 1.4375,
+      "learning_rate": 9.325256364581403e-06,
+      "loss": 0.022004596889019012,
+      "mean_token_accuracy": 0.9910150520503521,
+      "num_tokens": 26586291.0,
+      "step": 5000
+    },
+    {
+      "epoch": 0.5360134003350083,
+      "eval_entropy": 0.020443486923235468,
+      "eval_loss": 0.01984340138733387,
+      "eval_mean_token_accuracy": 0.9918238368034363,
+      "eval_num_tokens": 26586291.0,
+      "eval_runtime": 20.253,
+      "eval_samples_per_second": 37.032,
+      "eval_steps_per_second": 18.516,
+      "step": 5000
+    },
+    {
+      "entropy": 0.02020404681409218,
+      "epoch": 0.5381574539363484,
+      "grad_norm": 2.546875,
+      "learning_rate": 9.255996294431251e-06,
+      "loss": 0.018390218913555145,
+      "mean_token_accuracy": 0.992999167367816,
+      "num_tokens": 26704783.0,
+      "step": 5020
+    },
+    {
+      "entropy": 0.01915527433520765,
+      "epoch": 0.5403015075376885,
+      "grad_norm": 2.078125,
+      "learning_rate": 9.186772094325482e-06,
+      "loss": 0.02050086110830307,
+      "mean_token_accuracy": 0.991879815980792,
+      "num_tokens": 26801816.0,
+      "step": 5040
+    },
+    {
+      "entropy": 0.01799158312455802,
+      "epoch": 0.5424455611390284,
+      "grad_norm": 1.828125,
+      "learning_rate": 9.117587101714046e-06,
+      "loss": 0.02155505567789078,
+      "mean_token_accuracy": 0.9908429838716983,
+      "num_tokens": 26908853.0,
+      "step": 5060
+    },
+    {
+      "entropy": 0.020375365777476874,
+      "epoch": 0.5445896147403685,
+      "grad_norm": 1.75,
+      "learning_rate": 9.048444652156607e-06,
+      "loss": 0.022749082744121553,
+      "mean_token_accuracy": 0.991591277718544,
+      "num_tokens": 27027467.0,
+      "step": 5080
+    },
+    {
+      "entropy": 0.01897250141009863,
+      "epoch": 0.5467336683417086,
+      "grad_norm": 1.9375,
+      "learning_rate": 8.979348079161753e-06,
+      "loss": 0.021764059364795686,
+      "mean_token_accuracy": 0.9915874369442463,
+      "num_tokens": 27140253.0,
+      "step": 5100
+    },
+    {
+      "entropy": 0.020416706607784364,
+      "epoch": 0.5488777219430486,
+      "grad_norm": 1.515625,
+      "learning_rate": 8.910300714026242e-06,
+      "loss": 0.019444821774959563,
+      "mean_token_accuracy": 0.9921912178397179,
+      "num_tokens": 27250990.0,
+      "step": 5120
+    },
+    {
+      "entropy": 0.019052070355746763,
+      "epoch": 0.5510217755443886,
+      "grad_norm": 0.79296875,
+      "learning_rate": 8.841305885674431e-06,
+      "loss": 0.0189539834856987,
+      "mean_token_accuracy": 0.994080551713705,
+      "num_tokens": 27355412.0,
+      "step": 5140
+    },
+    {
+      "entropy": 0.019072407830481097,
+      "epoch": 0.5531658291457286,
+      "grad_norm": 2.125,
+      "learning_rate": 8.772366920497754e-06,
+      "loss": 0.022655683755874633,
+      "mean_token_accuracy": 0.9920244321227074,
+      "num_tokens": 27451798.0,
+      "step": 5160
+    },
+    {
+      "entropy": 0.0204707554947845,
+      "epoch": 0.5553098827470687,
+      "grad_norm": 1.765625,
+      "learning_rate": 8.70348714219436e-06,
+      "loss": 0.02314187288284302,
+      "mean_token_accuracy": 0.9906669486314058,
+      "num_tokens": 27566715.0,
+      "step": 5180
+    },
+    {
+      "entropy": 0.01811999880219446,
+      "epoch": 0.5574539363484087,
+      "grad_norm": 1.375,
+      "learning_rate": 8.634669871608879e-06,
+      "loss": 0.021425962448120117,
+      "mean_token_accuracy": 0.9925322774797678,
+      "num_tokens": 27659771.0,
+      "step": 5200
+    },
+    {
+      "entropy": 0.02080239012641414,
+      "epoch": 0.5595979899497487,
+      "grad_norm": 1.921875,
+      "learning_rate": 8.565918426572287e-06,
+      "loss": 0.021234531700611115,
+      "mean_token_accuracy": 0.9914034195244312,
+      "num_tokens": 27774261.0,
+      "step": 5220
+    },
+    {
+      "entropy": 0.018487965166877986,
+      "epoch": 0.5617420435510888,
+      "grad_norm": 1.7421875,
+      "learning_rate": 8.497236121741988e-06,
+      "loss": 0.020122981071472167,
+      "mean_token_accuracy": 0.9918253514915705,
+      "num_tokens": 27873405.0,
+      "step": 5240
+    },
+    {
+      "entropy": 0.020079654125311208,
+      "epoch": 0.5638860971524288,
+      "grad_norm": 1.875,
+      "learning_rate": 8.42862626844198e-06,
+      "loss": 0.0222967728972435,
+      "mean_token_accuracy": 0.9918321210891008,
+      "num_tokens": 27980906.0,
+      "step": 5260
+    },
+    {
+      "entropy": 0.02020533189497655,
+      "epoch": 0.5660301507537688,
+      "grad_norm": 2.046875,
+      "learning_rate": 8.3600921745032e-06,
+      "loss": 0.02204359769821167,
+      "mean_token_accuracy": 0.9903446055948735,
+      "num_tokens": 28085101.0,
+      "step": 5280
+    },
+    {
+      "entropy": 0.019518600214541947,
+      "epoch": 0.5681742043551089,
+      "grad_norm": 1.140625,
+      "learning_rate": 8.291637144104084e-06,
+      "loss": 0.02309599071741104,
+      "mean_token_accuracy": 0.9912882562726736,
+      "num_tokens": 28193879.0,
+      "step": 5300
+    },
+    {
+      "entropy": 0.02015163512619438,
+      "epoch": 0.5703182579564489,
+      "grad_norm": 1.828125,
+      "learning_rate": 8.223264477611218e-06,
+      "loss": 0.020999400317668913,
+      "mean_token_accuracy": 0.9915294233709574,
+      "num_tokens": 28296254.0,
+      "step": 5320
+    },
+    {
+      "entropy": 0.02264108727913481,
+      "epoch": 0.5724623115577889,
+      "grad_norm": 1.609375,
+      "learning_rate": 8.154977471420266e-06,
+      "loss": 0.02454686313867569,
+      "mean_token_accuracy": 0.9913088951259852,
+      "num_tokens": 28409744.0,
+      "step": 5340
+    },
+    {
+      "entropy": 0.02133122851785174,
+      "epoch": 0.574606365159129,
+      "grad_norm": 1.078125,
+      "learning_rate": 8.086779417797005e-06,
+      "loss": 0.018969587981700897,
+      "mean_token_accuracy": 0.9918996881693601,
+      "num_tokens": 28523944.0,
+      "step": 5360
+    },
+    {
+      "entropy": 0.020475312656594725,
+      "epoch": 0.5767504187604691,
+      "grad_norm": 2.578125,
+      "learning_rate": 8.018673604718621e-06,
+      "loss": 0.020480218529701232,
+      "mean_token_accuracy": 0.9909398626536131,
+      "num_tokens": 28620491.0,
+      "step": 5380
+    },
+    {
+      "entropy": 0.02034863719095483,
+      "epoch": 0.578894472361809,
+      "grad_norm": 2.3125,
+      "learning_rate": 7.950663315715189e-06,
+      "loss": 0.020605732500553132,
+      "mean_token_accuracy": 0.991118798777461,
+      "num_tokens": 28724436.0,
+      "step": 5400
+    },
+    {
+      "entropy": 0.019775752684540748,
+      "epoch": 0.5810385259631491,
+      "grad_norm": 3.78125,
+      "learning_rate": 7.882751829711343e-06,
+      "loss": 0.017187023162841798,
+      "mean_token_accuracy": 0.9925138726830482,
+      "num_tokens": 28834817.0,
+      "step": 5420
+    },
+    {
+      "entropy": 0.020291093751802693,
+      "epoch": 0.5831825795644892,
+      "grad_norm": 2.59375,
+      "learning_rate": 7.814942420868225e-06,
+      "loss": 0.02155611515045166,
+      "mean_token_accuracy": 0.9925528045743703,
+      "num_tokens": 28949852.0,
+      "step": 5440
+    },
+    {
+      "entropy": 0.020088849540070442,
+      "epoch": 0.5853266331658291,
+      "grad_norm": 1.4765625,
+      "learning_rate": 7.747238358425603e-06,
+      "loss": 0.021509405970573426,
+      "mean_token_accuracy": 0.9916152823716402,
+      "num_tokens": 29053126.0,
+      "step": 5460
+    },
+    {
+      "entropy": 0.02078960820804241,
+      "epoch": 0.5874706867671692,
+      "grad_norm": 1.6953125,
+      "learning_rate": 7.679642906544268e-06,
+      "loss": 0.022637182474136354,
+      "mean_token_accuracy": 0.9910664647817612,
+      "num_tokens": 29173776.0,
+      "step": 5480
+    },
+    {
+      "entropy": 0.019441102636301368,
+      "epoch": 0.5896147403685092,
+      "grad_norm": 3.515625,
+      "learning_rate": 7.6121593241486644e-06,
+      "loss": 0.021338345110416414,
+      "mean_token_accuracy": 0.9917159400880337,
+      "num_tokens": 29278255.0,
+      "step": 5500
+    },
+    {
+      "epoch": 0.5896147403685092,
+      "eval_entropy": 0.019579512438397312,
+      "eval_loss": 0.019862476736307144,
+      "eval_mean_token_accuracy": 0.9916693598429362,
+      "eval_num_tokens": 29278255.0,
+      "eval_runtime": 20.2793,
+      "eval_samples_per_second": 36.984,
+      "eval_steps_per_second": 18.492,
+      "step": 5500
+    },
+    {
+      "entropy": 0.01876566273849676,
+      "epoch": 0.5917587939698492,
+      "grad_norm": 1.53125,
+      "learning_rate": 7.5447908647697506e-06,
+      "loss": 0.02028908282518387,
+      "mean_token_accuracy": 0.9926750350743532,
+      "num_tokens": 29377393.0,
+      "step": 5520
+    },
+    {
+      "entropy": 0.019120201245004863,
+      "epoch": 0.5939028475711893,
+      "grad_norm": 2.09375,
+      "learning_rate": 7.47754077638817e-06,
+      "loss": 0.02107948064804077,
+      "mean_token_accuracy": 0.9924666147679091,
+      "num_tokens": 29481829.0,
+      "step": 5540
+    },
+    {
+      "entropy": 0.019154090119650392,
+      "epoch": 0.5960469011725293,
+      "grad_norm": 1.65625,
+      "learning_rate": 7.410412301277628e-06,
+      "loss": 0.021442399919033052,
+      "mean_token_accuracy": 0.9915456175804138,
+      "num_tokens": 29577548.0,
+      "step": 5560
+    },
+    {
+      "entropy": 0.020240560936554176,
+      "epoch": 0.5981909547738693,
+      "grad_norm": 1.71875,
+      "learning_rate": 7.343408675848593e-06,
+      "loss": 0.022295641899108886,
+      "mean_token_accuracy": 0.9908768083900213,
+      "num_tokens": 29693831.0,
+      "step": 5580
+    },
+    {
+      "entropy": 0.018894769826965786,
+      "epoch": 0.6003350083752094,
+      "grad_norm": 2.203125,
+      "learning_rate": 7.27653313049226e-06,
+      "loss": 0.017608430981636048,
+      "mean_token_accuracy": 0.9934392455965281,
+      "num_tokens": 29807626.0,
+      "step": 5600
+    },
+    {
+      "entropy": 0.018964907280769695,
+      "epoch": 0.6024790619765494,
+      "grad_norm": 1.7890625,
+      "learning_rate": 7.209788889424797e-06,
+      "loss": 0.021428127586841584,
+      "mean_token_accuracy": 0.992251905053854,
+      "num_tokens": 29918699.0,
+      "step": 5620
+    },
+    {
+      "entropy": 0.019173358734906287,
+      "epoch": 0.6046231155778894,
+      "grad_norm": 2.03125,
+      "learning_rate": 7.143179170531917e-06,
+      "loss": 0.02228628993034363,
+      "mean_token_accuracy": 0.9917910866439342,
+      "num_tokens": 30012605.0,
+      "step": 5640
+    },
+    {
+      "entropy": 0.020879295714121327,
+      "epoch": 0.6067671691792295,
+      "grad_norm": 1.71875,
+      "learning_rate": 7.076707185213716e-06,
+      "loss": 0.023567219078540803,
+      "mean_token_accuracy": 0.9915827866643667,
+      "num_tokens": 30121162.0,
+      "step": 5660
+    },
+    {
+      "entropy": 0.021870423488599043,
+      "epoch": 0.6089112227805695,
+      "grad_norm": 2.09375,
+      "learning_rate": 7.010376138229848e-06,
+      "loss": 0.023693473637104036,
+      "mean_token_accuracy": 0.9895996507257223,
+      "num_tokens": 30224756.0,
+      "step": 5680
+    },
+    {
+      "entropy": 0.02007417911304401,
+      "epoch": 0.6110552763819096,
+      "grad_norm": 1.6484375,
+      "learning_rate": 6.944189227545035e-06,
+      "loss": 0.018479639291763307,
+      "mean_token_accuracy": 0.992578137665987,
+      "num_tokens": 30328568.0,
+      "step": 5700
+    },
+    {
+      "entropy": 0.02106366403361335,
+      "epoch": 0.6131993299832496,
+      "grad_norm": 3.3125,
+      "learning_rate": 6.878149644174868e-06,
+      "loss": 0.021388374269008636,
+      "mean_token_accuracy": 0.9912810981273651,
+      "num_tokens": 30433800.0,
+      "step": 5720
+    },
+    {
+      "entropy": 0.019824264124895307,
+      "epoch": 0.6153433835845896,
+      "grad_norm": 3.828125,
+      "learning_rate": 6.812260572031966e-06,
+      "loss": 0.019826412200927734,
+      "mean_token_accuracy": 0.9915285713970661,
+      "num_tokens": 30533876.0,
+      "step": 5740
+    },
+    {
+      "entropy": 0.019220966260763816,
+      "epoch": 0.6174874371859297,
+      "grad_norm": 2.3125,
+      "learning_rate": 6.74652518777247e-06,
+      "loss": 0.023332512378692626,
+      "mean_token_accuracy": 0.9904117193073034,
+      "num_tokens": 30640243.0,
+      "step": 5760
+    },
+    {
+      "entropy": 0.019488869791348406,
+      "epoch": 0.6196314907872696,
+      "grad_norm": 1.7734375,
+      "learning_rate": 6.680946660642893e-06,
+      "loss": 0.023433607816696168,
+      "mean_token_accuracy": 0.991485184058547,
+      "num_tokens": 30740058.0,
+      "step": 5780
+    },
+    {
+      "entropy": 0.02283499206282613,
+      "epoch": 0.6217755443886097,
+      "grad_norm": 1.9296875,
+      "learning_rate": 6.615528152327328e-06,
+      "loss": 0.024059185385704042,
+      "mean_token_accuracy": 0.9913460418581963,
+      "num_tokens": 30852207.0,
+      "step": 5800
+    },
+    {
+      "entropy": 0.020180349141332952,
+      "epoch": 0.6239195979899498,
+      "grad_norm": 3.109375,
+      "learning_rate": 6.550272816795006e-06,
+      "loss": 0.01820368319749832,
+      "mean_token_accuracy": 0.9933834221214056,
+      "num_tokens": 30953138.0,
+      "step": 5820
+    },
+    {
+      "entropy": 0.019724707261821094,
+      "epoch": 0.6260636515912897,
+      "grad_norm": 1.65625,
+      "learning_rate": 6.485183800148248e-06,
+      "loss": 0.020257289707660674,
+      "mean_token_accuracy": 0.9930203214287758,
+      "num_tokens": 31064272.0,
+      "step": 5840
+    },
+    {
+      "entropy": 0.01873356030444029,
+      "epoch": 0.6282077051926298,
+      "grad_norm": 3.484375,
+      "learning_rate": 6.420264240470771e-06,
+      "loss": 0.02272033840417862,
+      "mean_token_accuracy": 0.9918572433292866,
+      "num_tokens": 31172652.0,
+      "step": 5860
+    },
+    {
+      "entropy": 0.019832546763041137,
+      "epoch": 0.6303517587939699,
+      "grad_norm": 2.25,
+      "learning_rate": 6.355517267676398e-06,
+      "loss": 0.022037458419799805,
+      "mean_token_accuracy": 0.9914021987468005,
+      "num_tokens": 31285622.0,
+      "step": 5880
+    },
+    {
+      "entropy": 0.021016045037094956,
+      "epoch": 0.6324958123953098,
+      "grad_norm": 1.9375,
+      "learning_rate": 6.290946003358175e-06,
+      "loss": 0.02272486835718155,
+      "mean_token_accuracy": 0.9905358534306288,
+      "num_tokens": 31391456.0,
+      "step": 5900
+    },
+    {
+      "entropy": 0.018993245228944033,
+      "epoch": 0.6346398659966499,
+      "grad_norm": 2.515625,
+      "learning_rate": 6.226553560637847e-06,
+      "loss": 0.02007617652416229,
+      "mean_token_accuracy": 0.9922155175358057,
+      "num_tokens": 31496166.0,
+      "step": 5920
+    },
+    {
+      "entropy": 0.0195752263855411,
+      "epoch": 0.63678391959799,
+      "grad_norm": 2.125,
+      "learning_rate": 6.162343044015784e-06,
+      "loss": 0.02081257700920105,
+      "mean_token_accuracy": 0.9921386137604713,
+      "num_tokens": 31605934.0,
+      "step": 5940
+    },
+    {
+      "entropy": 0.019383694695488885,
+      "epoch": 0.63892797319933,
+      "grad_norm": 2.046875,
+      "learning_rate": 6.098317549221302e-06,
+      "loss": 0.02021864652633667,
+      "mean_token_accuracy": 0.9916181784123183,
+      "num_tokens": 31712938.0,
+      "step": 5960
+    },
+    {
+      "entropy": 0.01924167173308433,
+      "epoch": 0.64107202680067,
+      "grad_norm": 2.046875,
+      "learning_rate": 6.034480163063401e-06,
+      "loss": 0.018102942407131194,
+      "mean_token_accuracy": 0.9921923886984587,
+      "num_tokens": 31813225.0,
+      "step": 5980
+    },
+    {
+      "entropy": 0.018873273960889492,
+      "epoch": 0.6432160804020101,
+      "grad_norm": 1.5234375,
+      "learning_rate": 5.970833963281968e-06,
+      "loss": 0.018628720939159394,
+      "mean_token_accuracy": 0.9928794108331204,
+      "num_tokens": 31928881.0,
+      "step": 6000
+    },
+    {
+      "epoch": 0.6432160804020101,
+      "eval_entropy": 0.01981416783077293,
+      "eval_loss": 0.019560638815164566,
+      "eval_mean_token_accuracy": 0.9923082728385926,
+      "eval_num_tokens": 31928881.0,
+      "eval_runtime": 20.2811,
+      "eval_samples_per_second": 36.98,
+      "eval_steps_per_second": 18.49,
+      "step": 6000
+    },
+    {
+      "entropy": 0.020261389288043573,
+      "epoch": 0.6453601340033501,
+      "grad_norm": 1.3125,
+      "learning_rate": 5.9073820183993635e-06,
+      "loss": 0.022888460755348207,
+      "mean_token_accuracy": 0.9925485417246819,
+      "num_tokens": 32038408.0,
+      "step": 6020
+    },
+    {
+      "entropy": 0.02017652262293268,
+      "epoch": 0.6475041876046901,
+      "grad_norm": 1.6328125,
+      "learning_rate": 5.8441273875725e-06,
+      "loss": 0.020387469232082366,
+      "mean_token_accuracy": 0.9918923225253821,
+      "num_tokens": 32145003.0,
+      "step": 6040
+    },
+    {
+      "entropy": 0.01868033191105951,
+      "epoch": 0.6496482412060302,
+      "grad_norm": 3.375,
+      "learning_rate": 5.781073120445353e-06,
+      "loss": 0.02243368774652481,
+      "mean_token_accuracy": 0.9915057402104139,
+      "num_tokens": 32251639.0,
+      "step": 6060
+    },
+    {
+      "entropy": 0.017508955025914475,
+      "epoch": 0.6517922948073702,
+      "grad_norm": 2.140625,
+      "learning_rate": 5.71822225700192e-06,
+      "loss": 0.017190653085708617,
+      "mean_token_accuracy": 0.9927597284317017,
+      "num_tokens": 32352371.0,
+      "step": 6080
+    },
+    {
+      "entropy": 0.019111776090062448,
+      "epoch": 0.6539363484087102,
+      "grad_norm": 1.1328125,
+      "learning_rate": 5.655577827419661e-06,
+      "loss": 0.02137412428855896,
+      "mean_token_accuracy": 0.9919875215739011,
+      "num_tokens": 32454655.0,
+      "step": 6100
+    },
+    {
+      "entropy": 0.023027155749787198,
+      "epoch": 0.6560804020100502,
+      "grad_norm": 2.015625,
+      "learning_rate": 5.5931428519234145e-06,
+      "loss": 0.02286381125450134,
+      "mean_token_accuracy": 0.9909462627023459,
+      "num_tokens": 32559559.0,
+      "step": 6120
+    },
+    {
+      "entropy": 0.02029506681608382,
+      "epoch": 0.6582244556113903,
+      "grad_norm": 3.21875,
+      "learning_rate": 5.530920340639776e-06,
+      "loss": 0.02172601819038391,
+      "mean_token_accuracy": 0.99092488437891,
+      "num_tokens": 32658365.0,
+      "step": 6140
+    },
+    {
+      "entropy": 0.021070733480109995,
+      "epoch": 0.6603685092127303,
+      "grad_norm": 3.625,
+      "learning_rate": 5.468913293451979e-06,
+      "loss": 0.021482250094413756,
+      "mean_token_accuracy": 0.9916323244571685,
+      "num_tokens": 32764039.0,
+      "step": 6160
+    },
+    {
+      "entropy": 0.018422811868731515,
+      "epoch": 0.6625125628140703,
+      "grad_norm": 2.6875,
+      "learning_rate": 5.407124699855259e-06,
+      "loss": 0.01822958290576935,
+      "mean_token_accuracy": 0.9923275396227836,
+      "num_tokens": 32877996.0,
+      "step": 6180
+    },
+    {
+      "entropy": 0.01931439164168296,
+      "epoch": 0.6646566164154104,
+      "grad_norm": 2.296875,
+      "learning_rate": 5.34555753881273e-06,
+      "loss": 0.021021611988544464,
+      "mean_token_accuracy": 0.9911792669445276,
+      "num_tokens": 32986191.0,
+      "step": 6200
+    },
+    {
+      "entropy": 0.019900657032940215,
+      "epoch": 0.6668006700167505,
+      "grad_norm": 1.734375,
+      "learning_rate": 5.284214778611758e-06,
+      "loss": 0.021909794211387633,
+      "mean_token_accuracy": 0.9917173352092504,
+      "num_tokens": 33098196.0,
+      "step": 6220
+    },
+    {
+      "entropy": 0.022017163155032902,
+      "epoch": 0.6689447236180904,
+      "grad_norm": 2.5625,
+      "learning_rate": 5.2230993767208524e-06,
+      "loss": 0.02269018590450287,
+      "mean_token_accuracy": 0.9912259686738253,
+      "num_tokens": 33211358.0,
+      "step": 6240
+    },
+    {
+      "entropy": 0.019263985499833325,
+      "epoch": 0.6710887772194305,
+      "grad_norm": 1.5625,
+      "learning_rate": 5.162214279647082e-06,
+      "loss": 0.023519453406333924,
+      "mean_token_accuracy": 0.9909868344664574,
+      "num_tokens": 33314953.0,
+      "step": 6260
+    },
+    {
+      "entropy": 0.021732687100166004,
+      "epoch": 0.6732328308207706,
+      "grad_norm": 2.265625,
+      "learning_rate": 5.1015624227940195e-06,
+      "loss": 0.021917779743671418,
+      "mean_token_accuracy": 0.9901779539883137,
+      "num_tokens": 33428051.0,
+      "step": 6280
+    },
+    {
+      "entropy": 0.019478338875705957,
+      "epoch": 0.6753768844221105,
+      "grad_norm": 2.296875,
+      "learning_rate": 5.041146730320215e-06,
+      "loss": 0.01874311566352844,
+      "mean_token_accuracy": 0.9923384670168162,
+      "num_tokens": 33537514.0,
+      "step": 6300
+    },
+    {
+      "entropy": 0.018508305449995533,
+      "epoch": 0.6775209380234506,
+      "grad_norm": 3.125,
+      "learning_rate": 4.980970114998216e-06,
+      "loss": 0.021730564534664154,
+      "mean_token_accuracy": 0.9915897861123085,
+      "num_tokens": 33639522.0,
+      "step": 6320
+    },
+    {
+      "entropy": 0.019665157535450818,
+      "epoch": 0.6796649916247907,
+      "grad_norm": 2.625,
+      "learning_rate": 4.921035478074142e-06,
+      "loss": 0.020458599925041197,
+      "mean_token_accuracy": 0.9918417375534773,
+      "num_tokens": 33743819.0,
+      "step": 6340
+    },
+    {
+      "entropy": 0.020174525574702784,
+      "epoch": 0.6818090452261306,
+      "grad_norm": 8.5,
+      "learning_rate": 4.861345709127803e-06,
+      "loss": 0.019577798247337342,
+      "mean_token_accuracy": 0.9921597026288509,
+      "num_tokens": 33847282.0,
+      "step": 6360
+    },
+    {
+      "entropy": 0.020318039059475268,
+      "epoch": 0.6839530988274707,
+      "grad_norm": 1.953125,
+      "learning_rate": 4.801903685933389e-06,
+      "loss": 0.020683969557285308,
+      "mean_token_accuracy": 0.9917214840650559,
+      "num_tokens": 33953961.0,
+      "step": 6380
+    },
+    {
+      "entropy": 0.01895209533279285,
+      "epoch": 0.6860971524288108,
+      "grad_norm": 2.046875,
+      "learning_rate": 4.742712274320725e-06,
+      "loss": 0.020713077485561372,
+      "mean_token_accuracy": 0.9920912489295006,
+      "num_tokens": 34054466.0,
+      "step": 6400
+    },
+    {
+      "entropy": 0.021405872826358065,
+      "epoch": 0.6882412060301507,
+      "grad_norm": 2.015625,
+      "learning_rate": 4.683774328037108e-06,
+      "loss": 0.024052520096302033,
+      "mean_token_accuracy": 0.9902271404862404,
+      "num_tokens": 34162751.0,
+      "step": 6420
+    },
+    {
+      "entropy": 0.02056694905677432,
+      "epoch": 0.6903852596314908,
+      "grad_norm": 2.109375,
+      "learning_rate": 4.625092688609711e-06,
+      "loss": 0.021541555225849152,
+      "mean_token_accuracy": 0.9906902097165584,
+      "num_tokens": 34271073.0,
+      "step": 6440
+    },
+    {
+      "entropy": 0.01796177522041944,
+      "epoch": 0.6925293132328308,
+      "grad_norm": 1.421875,
+      "learning_rate": 4.566670185208599e-06,
+      "loss": 0.017521923780441283,
+      "mean_token_accuracy": 0.9920081544667483,
+      "num_tokens": 34379697.0,
+      "step": 6460
+    },
+    {
+      "entropy": 0.01806882056660015,
+      "epoch": 0.6946733668341708,
+      "grad_norm": 1.6953125,
+      "learning_rate": 4.5085096345103205e-06,
+      "loss": 0.018725714087486266,
+      "mean_token_accuracy": 0.9868147686123848,
+      "num_tokens": 34483763.0,
+      "step": 6480
+    },
+    {
+      "entropy": 0.019470379299946217,
+      "epoch": 0.6968174204355109,
+      "grad_norm": 1.9140625,
+      "learning_rate": 4.450613840562113e-06,
+      "loss": 0.017752458155155183,
+      "mean_token_accuracy": 0.9928225323557853,
+      "num_tokens": 34588293.0,
+      "step": 6500
+    },
+    {
+      "epoch": 0.6968174204355109,
+      "eval_entropy": 0.019607524218813826,
+      "eval_loss": 0.019470153376460075,
+      "eval_mean_token_accuracy": 0.9920525077184041,
+      "eval_num_tokens": 34588293.0,
+      "eval_runtime": 20.2964,
+      "eval_samples_per_second": 36.952,
+      "eval_steps_per_second": 18.476,
+      "step": 6500
+    },
+    {
+      "entropy": 0.01848742664128622,
+      "epoch": 0.6989614740368509,
+      "grad_norm": 2.671875,
+      "learning_rate": 4.392985594646711e-06,
+      "loss": 0.02162926346063614,
+      "mean_token_accuracy": 0.9913080882281065,
+      "num_tokens": 34689227.0,
+      "step": 6520
+    },
+    {
+      "entropy": 0.019617287777646198,
+      "epoch": 0.701105527638191,
+      "grad_norm": 1.4765625,
+      "learning_rate": 4.335627675147774e-06,
+      "loss": 0.02028166800737381,
+      "mean_token_accuracy": 0.9930087491869927,
+      "num_tokens": 34791268.0,
+      "step": 6540
+    },
+    {
+      "entropy": 0.01906899958651138,
+      "epoch": 0.703249581239531,
+      "grad_norm": 2.234375,
+      "learning_rate": 4.278542847415936e-06,
+      "loss": 0.019539786875247954,
+      "mean_token_accuracy": 0.9920954536646605,
+      "num_tokens": 34894826.0,
+      "step": 6560
+    },
+    {
+      "entropy": 0.0211970092474985,
+      "epoch": 0.705393634840871,
+      "grad_norm": 1.5,
+      "learning_rate": 4.221733863635479e-06,
+      "loss": 0.023706386983394622,
+      "mean_token_accuracy": 0.990295572206378,
+      "num_tokens": 35009299.0,
+      "step": 6580
+    },
+    {
+      "entropy": 0.021638639217917442,
+      "epoch": 0.7075376884422111,
+      "grad_norm": 1.59375,
+      "learning_rate": 4.165203462691645e-06,
+      "loss": 0.021413305401802064,
+      "mean_token_accuracy": 0.9913092136383057,
+      "num_tokens": 35116618.0,
+      "step": 6600
+    },
+    {
+      "entropy": 0.02009705953132652,
+      "epoch": 0.7096817420435511,
+      "grad_norm": 1.9296875,
+      "learning_rate": 4.1089543700385885e-06,
+      "loss": 0.023159290850162505,
+      "mean_token_accuracy": 0.9899264223873615,
+      "num_tokens": 35214619.0,
+      "step": 6620
+    },
+    {
+      "entropy": 0.020903292103525928,
+      "epoch": 0.7118257956448911,
+      "grad_norm": 2.8125,
+      "learning_rate": 4.0529892975679775e-06,
+      "loss": 0.02029138207435608,
+      "mean_token_accuracy": 0.9917101729661226,
+      "num_tokens": 35331847.0,
+      "step": 6640
+    },
+    {
+      "entropy": 0.019355950780254715,
+      "epoch": 0.7139698492462312,
+      "grad_norm": 2.25,
+      "learning_rate": 3.997310943478246e-06,
+      "loss": 0.022747300565242767,
+      "mean_token_accuracy": 0.9911560833454132,
+      "num_tokens": 35435144.0,
+      "step": 6660
+    },
+    {
+      "entropy": 0.019431080123217724,
+      "epoch": 0.7161139028475711,
+      "grad_norm": 2.46875,
+      "learning_rate": 3.941921992144509e-06,
+      "loss": 0.022220295667648316,
+      "mean_token_accuracy": 0.9914675705134869,
+      "num_tokens": 35540146.0,
+      "step": 6680
+    },
+    {
+      "entropy": 0.021516031174610362,
+      "epoch": 0.7182579564489112,
+      "grad_norm": 1.0078125,
+      "learning_rate": 3.886825113989142e-06,
+      "loss": 0.020635062456130983,
+      "mean_token_accuracy": 0.9917585879564286,
+      "num_tokens": 35652696.0,
+      "step": 6700
+    },
+    {
+      "entropy": 0.019729190898897287,
+      "epoch": 0.7204020100502513,
+      "grad_norm": 2.09375,
+      "learning_rate": 3.832022965353032e-06,
+      "loss": 0.019225403666496277,
+      "mean_token_accuracy": 0.9942506503313779,
+      "num_tokens": 35757381.0,
+      "step": 6720
+    },
+    {
+      "entropy": 0.020393153377153795,
+      "epoch": 0.7225460636515912,
+      "grad_norm": 5.96875,
+      "learning_rate": 3.777518188367517e-06,
+      "loss": 0.02378985732793808,
+      "mean_token_accuracy": 0.9912960890680551,
+      "num_tokens": 35865336.0,
+      "step": 6740
+    },
+    {
+      "entropy": 0.019717699302668736,
+      "epoch": 0.7246901172529313,
+      "grad_norm": 1.8515625,
+      "learning_rate": 3.723313410826992e-06,
+      "loss": 0.020520621538162233,
+      "mean_token_accuracy": 0.9927110608667136,
+      "num_tokens": 35969262.0,
+      "step": 6760
+    },
+    {
+      "entropy": 0.020156670222718276,
+      "epoch": 0.7268341708542714,
+      "grad_norm": 2.953125,
+      "learning_rate": 3.6694112460622267e-06,
+      "loss": 0.022173659503459932,
+      "mean_token_accuracy": 0.9913631960749626,
+      "num_tokens": 36078203.0,
+      "step": 6780
+    },
+    {
+      "entropy": 0.02017759628906788,
+      "epoch": 0.7289782244556114,
+      "grad_norm": 1.6328125,
+      "learning_rate": 3.615814292814367e-06,
+      "loss": 0.02243554890155792,
+      "mean_token_accuracy": 0.9910614918917418,
+      "num_tokens": 36179468.0,
+      "step": 6800
+    },
+    {
+      "entropy": 0.019291495129346005,
+      "epoch": 0.7311222780569514,
+      "grad_norm": 2.03125,
+      "learning_rate": 3.562525135109649e-06,
+      "loss": 0.021489365398883818,
+      "mean_token_accuracy": 0.992356463521719,
+      "num_tokens": 36287569.0,
+      "step": 6820
+    },
+    {
+      "entropy": 0.017877134253399162,
+      "epoch": 0.7332663316582915,
+      "grad_norm": 2.8125,
+      "learning_rate": 3.509546342134796e-06,
+      "loss": 0.01892261952161789,
+      "mean_token_accuracy": 0.9919790286570788,
+      "num_tokens": 36403262.0,
+      "step": 6840
+    },
+    {
+      "entropy": 0.01958747923711144,
+      "epoch": 0.7354103852596315,
+      "grad_norm": 3.3125,
+      "learning_rate": 3.4568804681131964e-06,
+      "loss": 0.019973862171173095,
+      "mean_token_accuracy": 0.9915773533284664,
+      "num_tokens": 36508584.0,
+      "step": 6860
+    },
+    {
+      "entropy": 0.019319427383879883,
+      "epoch": 0.7375544388609715,
+      "grad_norm": 2.296875,
+      "learning_rate": 3.4045300521817105e-06,
+      "loss": 0.02136710286140442,
+      "mean_token_accuracy": 0.9920724961906672,
+      "num_tokens": 36610038.0,
+      "step": 6880
+    },
+    {
+      "entropy": 0.020166188537677952,
+      "epoch": 0.7396984924623116,
+      "grad_norm": 1.9453125,
+      "learning_rate": 3.3524976182682843e-06,
+      "loss": 0.018660889565944673,
+      "mean_token_accuracy": 0.9929303977638483,
+      "num_tokens": 36708085.0,
+      "step": 6900
+    },
+    {
+      "entropy": 0.018660859293049725,
+      "epoch": 0.7418425460636516,
+      "grad_norm": 2.40625,
+      "learning_rate": 3.3007856749702573e-06,
+      "loss": 0.01911092400550842,
+      "mean_token_accuracy": 0.9929109118878842,
+      "num_tokens": 36809564.0,
+      "step": 6920
+    },
+    {
+      "entropy": 0.020304086872602056,
+      "epoch": 0.7439865996649916,
+      "grad_norm": 1.734375,
+      "learning_rate": 3.2493967154334016e-06,
+      "loss": 0.019811296463012697,
+      "mean_token_accuracy": 0.9924409098923206,
+      "num_tokens": 36912901.0,
+      "step": 6940
+    },
+    {
+      "entropy": 0.017767412251100723,
+      "epoch": 0.7461306532663317,
+      "grad_norm": 1.515625,
+      "learning_rate": 3.198333217231757e-06,
+      "loss": 0.016918018460273743,
+      "mean_token_accuracy": 0.9936570934951305,
+      "num_tokens": 37028495.0,
+      "step": 6960
+    },
+    {
+      "entropy": 0.01919960001034724,
+      "epoch": 0.7482747068676717,
+      "grad_norm": 1.78125,
+      "learning_rate": 3.1475976422481526e-06,
+      "loss": 0.018871791660785675,
+      "mean_token_accuracy": 0.9923384845256805,
+      "num_tokens": 37128754.0,
+      "step": 6980
+    },
+    {
+      "entropy": 0.01957385808059371,
+      "epoch": 0.7504187604690117,
+      "grad_norm": 2.515625,
+      "learning_rate": 3.0971924365555116e-06,
+      "loss": 0.018768243491649628,
+      "mean_token_accuracy": 0.9924201011657715,
+      "num_tokens": 37230511.0,
+      "step": 7000
+    },
+    {
+      "epoch": 0.7504187604690117,
+      "eval_entropy": 0.019512116880068788,
+      "eval_loss": 0.019492411985993385,
+      "eval_mean_token_accuracy": 0.9919790083567301,
+      "eval_num_tokens": 37230511.0,
+      "eval_runtime": 20.2777,
+      "eval_samples_per_second": 36.986,
+      "eval_steps_per_second": 18.493,
+      "step": 7000
+    },
+    {
+      "entropy": 0.018825849278618988,
+      "epoch": 0.7525628140703517,
+      "grad_norm": 2.265625,
+      "learning_rate": 3.0471200302989535e-06,
+      "loss": 0.020853203535079957,
+      "mean_token_accuracy": 0.9906294353306293,
+      "num_tokens": 37332200.0,
+      "step": 7020
+    },
+    {
+      "entropy": 0.019942389463130893,
+      "epoch": 0.7547068676716918,
+      "grad_norm": 2.203125,
+      "learning_rate": 2.9973828375785864e-06,
+      "loss": 0.018565280735492705,
+      "mean_token_accuracy": 0.9909187290817499,
+      "num_tokens": 37433333.0,
+      "step": 7040
+    },
+    {
+      "entropy": 0.020680780549423617,
+      "epoch": 0.7568509212730318,
+      "grad_norm": 1.3984375,
+      "learning_rate": 2.947983256333166e-06,
+      "loss": 0.022726891934871672,
+      "mean_token_accuracy": 0.9920847479254007,
+      "num_tokens": 37535155.0,
+      "step": 7060
+    },
+    {
+      "entropy": 0.01799933227546262,
+      "epoch": 0.7589949748743718,
+      "grad_norm": 2.375,
+      "learning_rate": 2.8989236682244372e-06,
+      "loss": 0.019517210125923157,
+      "mean_token_accuracy": 0.9922154374420643,
+      "num_tokens": 37642429.0,
+      "step": 7080
+    },
+    {
+      "entropy": 0.018576059823135437,
+      "epoch": 0.7611390284757119,
+      "grad_norm": 1.8671875,
+      "learning_rate": 2.8502064385223437e-06,
+      "loss": 0.02000274509191513,
+      "mean_token_accuracy": 0.9927195519208908,
+      "num_tokens": 37743382.0,
+      "step": 7100
+    },
+    {
+      "entropy": 0.01817750639111182,
+      "epoch": 0.763283082077052,
+      "grad_norm": 0.76953125,
+      "learning_rate": 2.8018339159909937e-06,
+      "loss": 0.018051585555076598,
+      "mean_token_accuracy": 0.9927114237099885,
+      "num_tokens": 37849802.0,
+      "step": 7120
+    },
+    {
+      "entropy": 0.01861468399677051,
+      "epoch": 0.7654271356783919,
+      "grad_norm": 2.453125,
+      "learning_rate": 2.7538084327753823e-06,
+      "loss": 0.020044369995594023,
+      "mean_token_accuracy": 0.9925104282796383,
+      "num_tokens": 37949510.0,
+      "step": 7140
+    },
+    {
+      "entropy": 0.01957492132596599,
+      "epoch": 0.767571189279732,
+      "grad_norm": 1.6640625,
+      "learning_rate": 2.706132304289012e-06,
+      "loss": 0.02030598223209381,
+      "mean_token_accuracy": 0.9924644228070975,
+      "num_tokens": 38056600.0,
+      "step": 7160
+    },
+    {
+      "entropy": 0.019891365615512767,
+      "epoch": 0.7697152428810721,
+      "grad_norm": 1.3046875,
+      "learning_rate": 2.6588078291022045e-06,
+      "loss": 0.020326973497867586,
+      "mean_token_accuracy": 0.9918508928269147,
+      "num_tokens": 38168190.0,
+      "step": 7180
+    },
+    {
+      "entropy": 0.01919556673688021,
+      "epoch": 0.771859296482412,
+      "grad_norm": 2.796875,
+      "learning_rate": 2.611837288831319e-06,
+      "loss": 0.02155778855085373,
+      "mean_token_accuracy": 0.9919852402061224,
+      "num_tokens": 38276246.0,
+      "step": 7200
+    },
+    {
+      "entropy": 0.019413144076725074,
+      "epoch": 0.7740033500837521,
+      "grad_norm": 5.8125,
+      "learning_rate": 2.565222948028746e-06,
+      "loss": 0.0233286514878273,
+      "mean_token_accuracy": 0.9901302725076675,
+      "num_tokens": 38380173.0,
+      "step": 7220
+    },
+    {
+      "entropy": 0.019048391740625448,
+      "epoch": 0.7761474036850922,
+      "grad_norm": 2.484375,
+      "learning_rate": 2.518967054073702e-06,
+      "loss": 0.01727575808763504,
+      "mean_token_accuracy": 0.9933784183114767,
+      "num_tokens": 38485326.0,
+      "step": 7240
+    },
+    {
+      "entropy": 0.018992443338993324,
+      "epoch": 0.7782914572864321,
+      "grad_norm": 1.765625,
+      "learning_rate": 2.473071837063921e-06,
+      "loss": 0.020196519792079926,
+      "mean_token_accuracy": 0.9929014749825,
+      "num_tokens": 38593033.0,
+      "step": 7260
+    },
+    {
+      "entropy": 0.01934261324011004,
+      "epoch": 0.7804355108877722,
+      "grad_norm": 2.078125,
+      "learning_rate": 2.427539509708092e-06,
+      "loss": 0.021069280803203583,
+      "mean_token_accuracy": 0.9920600451529026,
+      "num_tokens": 38695209.0,
+      "step": 7280
+    },
+    {
+      "entropy": 0.01937621497809232,
+      "epoch": 0.7825795644891123,
+      "grad_norm": 1.96875,
+      "learning_rate": 2.382372267219212e-06,
+      "loss": 0.02313290983438492,
+      "mean_token_accuracy": 0.9915325742214918,
+      "num_tokens": 38804750.0,
+      "step": 7300
+    },
+    {
+      "entropy": 0.02088976109387204,
+      "epoch": 0.7847236180904522,
+      "grad_norm": 3.390625,
+      "learning_rate": 2.3375722872087435e-06,
+      "loss": 0.02172703593969345,
+      "mean_token_accuracy": 0.9911188166588545,
+      "num_tokens": 38909561.0,
+      "step": 7320
+    },
+    {
+      "entropy": 0.019513376194367993,
+      "epoch": 0.7868676716917923,
+      "grad_norm": 1.6171875,
+      "learning_rate": 2.293141729581606e-06,
+      "loss": 0.020942328870296477,
+      "mean_token_accuracy": 0.9914060894399881,
+      "num_tokens": 39024348.0,
+      "step": 7340
+    },
+    {
+      "entropy": 0.019013147754321835,
+      "epoch": 0.7890117252931323,
+      "grad_norm": 3.03125,
+      "learning_rate": 2.24908273643208e-06,
+      "loss": 0.01895921677350998,
+      "mean_token_accuracy": 0.9920435093343258,
+      "num_tokens": 39123358.0,
+      "step": 7360
+    },
+    {
+      "entropy": 0.01850612224473025,
+      "epoch": 0.7911557788944724,
+      "grad_norm": 2.0625,
+      "learning_rate": 2.20539743194049e-06,
+      "loss": 0.020412975549697877,
+      "mean_token_accuracy": 0.992505956813693,
+      "num_tokens": 39222173.0,
+      "step": 7380
+    },
+    {
+      "entropy": 0.018990432533109926,
+      "epoch": 0.7932998324958124,
+      "grad_norm": 2.515625,
+      "learning_rate": 2.162087922270827e-06,
+      "loss": 0.019987034797668456,
+      "mean_token_accuracy": 0.9921554632484912,
+      "num_tokens": 39324555.0,
+      "step": 7400
+    },
+    {
+      "entropy": 0.020331120665650813,
+      "epoch": 0.7954438860971524,
+      "grad_norm": 2.59375,
+      "learning_rate": 2.119156295469188e-06,
+      "loss": 0.020988868176937105,
+      "mean_token_accuracy": 0.992167953774333,
+      "num_tokens": 39436930.0,
+      "step": 7420
+    },
+    {
+      "entropy": 0.01908351750444126,
+      "epoch": 0.7975879396984925,
+      "grad_norm": 2.140625,
+      "learning_rate": 2.0766046213631085e-06,
+      "loss": 0.02441962659358978,
+      "mean_token_accuracy": 0.9905336666852236,
+      "num_tokens": 39539754.0,
+      "step": 7440
+    },
+    {
+      "entropy": 0.01844053882264234,
+      "epoch": 0.7997319932998325,
+      "grad_norm": 1.4609375,
+      "learning_rate": 2.034434951461787e-06,
+      "loss": 0.021246211230754854,
+      "mean_token_accuracy": 0.9916405260562897,
+      "num_tokens": 39642853.0,
+      "step": 7460
+    },
+    {
+      "entropy": 0.019568845091521327,
+      "epoch": 0.8018760469011725,
+      "grad_norm": 2.234375,
+      "learning_rate": 1.992649318857145e-06,
+      "loss": 0.02266250252723694,
+      "mean_token_accuracy": 0.9901510171592236,
+      "num_tokens": 39753649.0,
+      "step": 7480
+    },
+    {
+      "entropy": 0.019649790661560473,
+      "epoch": 0.8040201005025126,
+      "grad_norm": 1.4296875,
+      "learning_rate": 1.951249738125841e-06,
+      "loss": 0.02106764316558838,
+      "mean_token_accuracy": 0.992297976091504,
+      "num_tokens": 39867091.0,
+      "step": 7500
+    },
+    {
+      "epoch": 0.8040201005025126,
+      "eval_entropy": 0.01951531241268579,
+      "eval_loss": 0.019423136487603188,
+      "eval_mean_token_accuracy": 0.9921902815500895,
+      "eval_num_tokens": 39867091.0,
+      "eval_runtime": 20.2653,
+      "eval_samples_per_second": 37.009,
+      "eval_steps_per_second": 18.504,
+      "step": 7500
+    },
+    {
+      "entropy": 0.01873672366637038,
+      "epoch": 0.8061641541038526,
+      "grad_norm": 2.515625,
+      "learning_rate": 1.9102382052321247e-06,
+      "loss": 0.02217361479997635,
+      "mean_token_accuracy": 0.991604308411479,
+      "num_tokens": 39975152.0,
+      "step": 7520
+    },
+    {
+      "entropy": 0.018056424487326694,
+      "epoch": 0.8083082077051926,
+      "grad_norm": 2.1875,
+      "learning_rate": 1.8696166974316098e-06,
+      "loss": 0.018713992834091187,
+      "mean_token_accuracy": 0.9923322558403015,
+      "num_tokens": 40073850.0,
+      "step": 7540
+    },
+    {
+      "entropy": 0.019656274415865482,
+      "epoch": 0.8104522613065327,
+      "grad_norm": 1.75,
+      "learning_rate": 1.829387173175957e-06,
+      "loss": 0.02295905351638794,
+      "mean_token_accuracy": 0.9911978255957365,
+      "num_tokens": 40174029.0,
+      "step": 7560
+    },
+    {
+      "entropy": 0.018212407533337683,
+      "epoch": 0.8125963149078727,
+      "grad_norm": 2.203125,
+      "learning_rate": 1.7895515720184298e-06,
+      "loss": 0.01762194335460663,
+      "mean_token_accuracy": 0.993547822907567,
+      "num_tokens": 40277338.0,
+      "step": 7580
+    },
+    {
+      "entropy": 0.019111268576693874,
+      "epoch": 0.8147403685092127,
+      "grad_norm": 3.078125,
+      "learning_rate": 1.7501118145204078e-06,
+      "loss": 0.02163155972957611,
+      "mean_token_accuracy": 0.9915014613419771,
+      "num_tokens": 40384919.0,
+      "step": 7600
+    },
+    {
+      "entropy": 0.020142577382034687,
+      "epoch": 0.8168844221105528,
+      "grad_norm": 2.1875,
+      "learning_rate": 1.71106980215878e-06,
+      "loss": 0.01972368061542511,
+      "mean_token_accuracy": 0.9922790337353945,
+      "num_tokens": 40491708.0,
+      "step": 7620
+    },
+    {
+      "entropy": 0.019609718415108546,
+      "epoch": 0.8190284757118927,
+      "grad_norm": 6.25,
+      "learning_rate": 1.6724274172342725e-06,
+      "loss": 0.023548129200935363,
+      "mean_token_accuracy": 0.991268851980567,
+      "num_tokens": 40604052.0,
+      "step": 7640
+    },
+    {
+      "entropy": 0.019134915648101013,
+      "epoch": 0.8211725293132328,
+      "grad_norm": 1.75,
+      "learning_rate": 1.6341865227807075e-06,
+      "loss": 0.019575177133083342,
+      "mean_token_accuracy": 0.992620923742652,
+      "num_tokens": 40712926.0,
+      "step": 7660
+    },
+    {
+      "entropy": 0.018662033730379335,
+      "epoch": 0.8233165829145729,
+      "grad_norm": 2.265625,
+      "learning_rate": 1.5963489624751595e-06,
+      "loss": 0.019268378615379333,
+      "mean_token_accuracy": 0.992533054947853,
+      "num_tokens": 40818181.0,
+      "step": 7680
+    },
+    {
+      "entropy": 0.019738298160768865,
+      "epoch": 0.825460636515913,
+      "grad_norm": 1.9921875,
+      "learning_rate": 1.5589165605490918e-06,
+      "loss": 0.018965110182762146,
+      "mean_token_accuracy": 0.9918511062860489,
+      "num_tokens": 40925382.0,
+      "step": 7700
+    },
+    {
+      "entropy": 0.01947517461164807,
+      "epoch": 0.8276046901172529,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.5218911217003985e-06,
+      "loss": 0.02183605879545212,
+      "mean_token_accuracy": 0.9911946535110474,
+      "num_tokens": 41044722.0,
+      "step": 7720
+    },
+    {
+      "entropy": 0.019855259767928145,
+      "epoch": 0.829748743718593,
+      "grad_norm": 1.90625,
+      "learning_rate": 1.4852744310063893e-06,
+      "loss": 0.020095731317996978,
+      "mean_token_accuracy": 0.9914216987788678,
+      "num_tokens": 41163232.0,
+      "step": 7740
+    },
+    {
+      "entropy": 0.01931553952094873,
+      "epoch": 0.831892797319933,
+      "grad_norm": 2.703125,
+      "learning_rate": 1.4490682538377344e-06,
+      "loss": 0.02068305015563965,
+      "mean_token_accuracy": 0.9912440560758113,
+      "num_tokens": 41275818.0,
+      "step": 7760
+    },
+    {
+      "entropy": 0.01949581656031114,
+      "epoch": 0.834036850921273,
+      "grad_norm": 2.640625,
+      "learning_rate": 1.4132743357733502e-06,
+      "loss": 0.019521139562129974,
+      "mean_token_accuracy": 0.9927113294601441,
+      "num_tokens": 41378446.0,
+      "step": 7780
+    },
+    {
+      "entropy": 0.019279925492878647,
+      "epoch": 0.8361809045226131,
+      "grad_norm": 1.71875,
+      "learning_rate": 1.3778944025162399e-06,
+      "loss": 0.020213739573955537,
+      "mean_token_accuracy": 0.9918138355016708,
+      "num_tokens": 41480586.0,
+      "step": 7800
+    },
+    {
+      "entropy": 0.01848278012148512,
+      "epoch": 0.8383249581239531,
+      "grad_norm": 1.8046875,
+      "learning_rate": 1.3429301598102973e-06,
+      "loss": 0.021417385339736937,
+      "mean_token_accuracy": 0.992524030432105,
+      "num_tokens": 41583797.0,
+      "step": 7820
+    },
+    {
+      "entropy": 0.019653417706990695,
+      "epoch": 0.8404690117252931,
+      "grad_norm": 1.4375,
+      "learning_rate": 1.3083832933580608e-06,
+      "loss": 0.019776181876659395,
+      "mean_token_accuracy": 0.9927833370864392,
+      "num_tokens": 41693464.0,
+      "step": 7840
+    },
+    {
+      "entropy": 0.019763251290942207,
+      "epoch": 0.8426130653266332,
+      "grad_norm": 3.109375,
+      "learning_rate": 1.2742554687394537e-06,
+      "loss": 0.02297937124967575,
+      "mean_token_accuracy": 0.992657694220543,
+      "num_tokens": 41790157.0,
+      "step": 7860
+    },
+    {
+      "entropy": 0.019936064798548613,
+      "epoch": 0.8447571189279732,
+      "grad_norm": 3.609375,
+      "learning_rate": 1.2405483313314738e-06,
+      "loss": 0.02262817770242691,
+      "mean_token_accuracy": 0.9905056063085794,
+      "num_tokens": 41893582.0,
+      "step": 7880
+    },
+    {
+      "entropy": 0.01934517203821997,
+      "epoch": 0.8469011725293132,
+      "grad_norm": 2.109375,
+      "learning_rate": 1.2072635062288673e-06,
+      "loss": 0.020036940276622773,
+      "mean_token_accuracy": 0.9923514679074288,
+      "num_tokens": 42004717.0,
+      "step": 7900
+    },
+    {
+      "entropy": 0.01836295023481398,
+      "epoch": 0.8490452261306533,
+      "grad_norm": 3.1875,
+      "learning_rate": 1.174402598165787e-06,
+      "loss": 0.017996013164520264,
+      "mean_token_accuracy": 0.9931914895772934,
+      "num_tokens": 42114536.0,
+      "step": 7920
+    },
+    {
+      "entropy": 0.020249022626080658,
+      "epoch": 0.8511892797319933,
+      "grad_norm": 2.5,
+      "learning_rate": 1.1419671914384134e-06,
+      "loss": 0.02248285710811615,
+      "mean_token_accuracy": 0.991206606850028,
+      "num_tokens": 42222812.0,
+      "step": 7940
+    },
+    {
+      "entropy": 0.02073315970085332,
+      "epoch": 0.8533333333333334,
+      "grad_norm": 3.0,
+      "learning_rate": 1.1099588498285806e-06,
+      "loss": 0.02276255041360855,
+      "mean_token_accuracy": 0.9906704191118478,
+      "num_tokens": 42337872.0,
+      "step": 7960
+    },
+    {
+      "entropy": 0.019947545273566902,
+      "epoch": 0.8554773869346733,
+      "grad_norm": 1.5390625,
+      "learning_rate": 1.0783791165283798e-06,
+      "loss": 0.0205001100897789,
+      "mean_token_accuracy": 0.9920544352382421,
+      "num_tokens": 42443680.0,
+      "step": 7980
+    },
+    {
+      "entropy": 0.017547584273529537,
+      "epoch": 0.8576214405360134,
+      "grad_norm": 3.328125,
+      "learning_rate": 1.0472295140657584e-06,
+      "loss": 0.01985829621553421,
+      "mean_token_accuracy": 0.9921385683119297,
+      "num_tokens": 42548237.0,
+      "step": 8000
+    },
+    {
+      "epoch": 0.8576214405360134,
+      "eval_entropy": 0.019516329319584958,
+      "eval_loss": 0.01938561163842678,
+      "eval_mean_token_accuracy": 0.9923241523106893,
+      "eval_num_tokens": 42548237.0,
+      "eval_runtime": 20.235,
+      "eval_samples_per_second": 37.064,
+      "eval_steps_per_second": 18.532,
+      "step": 8000
+    },
+    {
+      "entropy": 0.018809675735496967,
+      "epoch": 0.8597654941373535,
+      "grad_norm": 1.5859375,
+      "learning_rate": 1.0165115442311168e-06,
+      "loss": 0.02014607787132263,
+      "mean_token_accuracy": 0.9919284801930189,
+      "num_tokens": 42659432.0,
+      "step": 8020
+    },
+    {
+      "entropy": 0.01733278074198097,
+      "epoch": 0.8619095477386934,
+      "grad_norm": 1.921875,
+      "learning_rate": 9.862266880049033e-07,
+      "loss": 0.018754176795482635,
+      "mean_token_accuracy": 0.9923460192978382,
+      "num_tokens": 42759466.0,
+      "step": 8040
+    },
+    {
+      "entropy": 0.01832941572265554,
+      "epoch": 0.8640536013400335,
+      "grad_norm": 2.078125,
+      "learning_rate": 9.563764054862112e-07,
+      "loss": 0.020297129452228547,
+      "mean_token_accuracy": 0.9928287036716938,
+      "num_tokens": 42868393.0,
+      "step": 8060
+    },
+    {
+      "entropy": 0.020013672889513147,
+      "epoch": 0.8661976549413736,
+      "grad_norm": 2.296875,
+      "learning_rate": 9.26962135822389e-07,
+      "loss": 0.021075569093227386,
+      "mean_token_accuracy": 0.9927308529615402,
+      "num_tokens": 42978572.0,
+      "step": 8080
+    },
+    {
+      "entropy": 0.017360182798279312,
+      "epoch": 0.8683417085427135,
+      "grad_norm": 1.5625,
+      "learning_rate": 8.979852971396508e-07,
+      "loss": 0.018198615312576293,
+      "mean_token_accuracy": 0.993445823341608,
+      "num_tokens": 43086831.0,
+      "step": 8100
+    },
+    {
+      "entropy": 0.02018861417880089,
+      "epoch": 0.8704857621440536,
+      "grad_norm": 2.25,
+      "learning_rate": 8.69447286474705e-07,
+      "loss": 0.02233450710773468,
+      "mean_token_accuracy": 0.9902903474867344,
+      "num_tokens": 43195060.0,
+      "step": 8120
+    },
+    {
+      "entropy": 0.019654828006423486,
+      "epoch": 0.8726298157453937,
+      "grad_norm": 2.515625,
+      "learning_rate": 8.413494797074095e-07,
+      "loss": 0.01920991241931915,
+      "mean_token_accuracy": 0.9916286468505859,
+      "num_tokens": 43306019.0,
+      "step": 8140
+    },
+    {
+      "entropy": 0.02208902873887837,
+      "epoch": 0.8747738693467336,
+      "grad_norm": 2.140625,
+      "learning_rate": 8.136932314944179e-07,
+      "loss": 0.022708958387374877,
+      "mean_token_accuracy": 0.9901477582752705,
+      "num_tokens": 43416895.0,
+      "step": 8160
+    },
+    {
+      "entropy": 0.01981349259772287,
+      "epoch": 0.8769179229480737,
+      "grad_norm": 2.71875,
+      "learning_rate": 7.864798752038949e-07,
+      "loss": 0.01928192377090454,
+      "mean_token_accuracy": 0.9919186655431986,
+      "num_tokens": 43516044.0,
+      "step": 8180
+    },
+    {
+      "entropy": 0.01925912519070607,
+      "epoch": 0.8790619765494138,
+      "grad_norm": 1.734375,
+      "learning_rate": 7.597107228512101e-07,
+      "loss": 0.02041231542825699,
+      "mean_token_accuracy": 0.9925541203469038,
+      "num_tokens": 43622616.0,
+      "step": 8200
+    },
+    {
+      "entropy": 0.01794401331856079,
+      "epoch": 0.8812060301507537,
+      "grad_norm": 1.40625,
+      "learning_rate": 7.333870650356911e-07,
+      "loss": 0.01903040409088135,
+      "mean_token_accuracy": 0.9926789086312056,
+      "num_tokens": 43738048.0,
+      "step": 8220
+    },
+    {
+      "entropy": 0.020185897200371984,
+      "epoch": 0.8833500837520938,
+      "grad_norm": 3.078125,
+      "learning_rate": 7.075101708784038e-07,
+      "loss": 0.02289024293422699,
+      "mean_token_accuracy": 0.9907027028501034,
+      "num_tokens": 43849936.0,
+      "step": 8240
+    },
+    {
+      "entropy": 0.020925003469278634,
+      "epoch": 0.8854941373534339,
+      "grad_norm": 2.96875,
+      "learning_rate": 6.820812879609551e-07,
+      "loss": 0.02114592641592026,
+      "mean_token_accuracy": 0.9914819072932005,
+      "num_tokens": 43953914.0,
+      "step": 8260
+    },
+    {
+      "entropy": 0.019342410219724115,
+      "epoch": 0.8876381909547739,
+      "grad_norm": 3.609375,
+      "learning_rate": 6.571016422653609e-07,
+      "loss": 0.020218712091445924,
+      "mean_token_accuracy": 0.9917961444705725,
+      "num_tokens": 44060574.0,
+      "step": 8280
+    },
+    {
+      "entropy": 0.018030824407537692,
+      "epoch": 0.8897822445561139,
+      "grad_norm": 1.171875,
+      "learning_rate": 6.325724381149201e-07,
+      "loss": 0.01951337456703186,
+      "mean_token_accuracy": 0.9863187871873379,
+      "num_tokens": 44169435.0,
+      "step": 8300
+    },
+    {
+      "entropy": 0.01994664284213741,
+      "epoch": 0.891926298157454,
+      "grad_norm": 2.921875,
+      "learning_rate": 6.084948581161643e-07,
+      "loss": 0.019201353192329407,
+      "mean_token_accuracy": 0.9925763100385666,
+      "num_tokens": 44276020.0,
+      "step": 8320
+    },
+    {
+      "entropy": 0.02081800427977214,
+      "epoch": 0.894070351758794,
+      "grad_norm": 1.6953125,
+      "learning_rate": 5.848700631018411e-07,
+      "loss": 0.020571966469287873,
+      "mean_token_accuracy": 0.9917529292404652,
+      "num_tokens": 44386885.0,
+      "step": 8340
+    },
+    {
+      "entropy": 0.01815503019108746,
+      "epoch": 0.896214405360134,
+      "grad_norm": 2.203125,
+      "learning_rate": 5.616991920749349e-07,
+      "loss": 0.020456781983375548,
+      "mean_token_accuracy": 0.9919084422290325,
+      "num_tokens": 44493234.0,
+      "step": 8360
+    },
+    {
+      "entropy": 0.018217097373872095,
+      "epoch": 0.898358458961474,
+      "grad_norm": 1.375,
+      "learning_rate": 5.389833621537755e-07,
+      "loss": 0.01739330738782883,
+      "mean_token_accuracy": 0.9934280533343554,
+      "num_tokens": 44596525.0,
+      "step": 8380
+    },
+    {
+      "entropy": 0.01944470095006636,
+      "epoch": 0.9005025125628141,
+      "grad_norm": 1.875,
+      "learning_rate": 5.167236685181532e-07,
+      "loss": 0.01959017515182495,
+      "mean_token_accuracy": 0.9926484901458025,
+      "num_tokens": 44704672.0,
+      "step": 8400
+    },
+    {
+      "entropy": 0.01922774040344848,
+      "epoch": 0.9026465661641541,
+      "grad_norm": 1.8671875,
+      "learning_rate": 4.949211843565449e-07,
+      "loss": 0.018985719978809358,
+      "mean_token_accuracy": 0.9930604692548514,
+      "num_tokens": 44802160.0,
+      "step": 8420
+    },
+    {
+      "entropy": 0.019086347535608183,
+      "epoch": 0.9047906197654941,
+      "grad_norm": 2.625,
+      "learning_rate": 4.7357696081435036e-07,
+      "loss": 0.021893636882305147,
+      "mean_token_accuracy": 0.9909486912190915,
+      "num_tokens": 44907874.0,
+      "step": 8440
+    },
+    {
+      "entropy": 0.01974412333381679,
+      "epoch": 0.9069346733668342,
+      "grad_norm": 3.1875,
+      "learning_rate": 4.5269202694322334e-07,
+      "loss": 0.0225388765335083,
+      "mean_token_accuracy": 0.9913734450936318,
+      "num_tokens": 45011015.0,
+      "step": 8460
+    },
+    {
+      "entropy": 0.0196786193088883,
+      "epoch": 0.9090787269681742,
+      "grad_norm": 3.484375,
+      "learning_rate": 4.3226738965146733e-07,
+      "loss": 0.019231879711151124,
+      "mean_token_accuracy": 0.9921259760856629,
+      "num_tokens": 45122211.0,
+      "step": 8480
+    },
+    {
+      "entropy": 0.017996370965624918,
+      "epoch": 0.9112227805695142,
+      "grad_norm": 1.3125,
+      "learning_rate": 4.1230403365547065e-07,
+      "loss": 0.018080990016460418,
+      "mean_token_accuracy": 0.9934276740998029,
+      "num_tokens": 45229362.0,
+      "step": 8500
+    },
+    {
+      "epoch": 0.9112227805695142,
+      "eval_entropy": 0.019543887719133637,
+      "eval_loss": 0.019348351284861565,
+      "eval_mean_token_accuracy": 0.9922665920257568,
+      "eval_num_tokens": 45229362.0,
+      "eval_runtime": 20.2762,
+      "eval_samples_per_second": 36.989,
+      "eval_steps_per_second": 18.495,
+      "step": 8500
+    },
+    {
+      "entropy": 0.021203386071078967,
+      "epoch": 0.9133668341708543,
+      "grad_norm": 1.609375,
+      "learning_rate": 3.9280292143225264e-07,
+      "loss": 0.022594137489795683,
+      "mean_token_accuracy": 0.991113456711173,
+      "num_tokens": 45338204.0,
+      "step": 8520
+    },
+    {
+      "entropy": 0.018270298717538935,
+      "epoch": 0.9155108877721944,
+      "grad_norm": 2.03125,
+      "learning_rate": 3.737649931730458e-07,
+      "loss": 0.019554194808006287,
+      "mean_token_accuracy": 0.9928128100931645,
+      "num_tokens": 45451611.0,
+      "step": 8540
+    },
+    {
+      "entropy": 0.01802048790036679,
+      "epoch": 0.9176549413735343,
+      "grad_norm": 1.8984375,
+      "learning_rate": 3.5519116673796707e-07,
+      "loss": 0.01832857131958008,
+      "mean_token_accuracy": 0.9920665070414543,
+      "num_tokens": 45552742.0,
+      "step": 8560
+    },
+    {
+      "entropy": 0.019545484634431887,
+      "epoch": 0.9197989949748744,
+      "grad_norm": 2.078125,
+      "learning_rate": 3.3708233761177935e-07,
+      "loss": 0.019848516583442687,
+      "mean_token_accuracy": 0.9930632192641496,
+      "num_tokens": 45654718.0,
+      "step": 8580
+    },
+    {
+      "entropy": 0.019058815206676627,
+      "epoch": 0.9219430485762145,
+      "grad_norm": 2.21875,
+      "learning_rate": 3.1943937886070197e-07,
+      "loss": 0.019996930658817292,
+      "mean_token_accuracy": 0.9920886129140853,
+      "num_tokens": 45764177.0,
+      "step": 8600
+    },
+    {
+      "entropy": 0.019873253857394958,
+      "epoch": 0.9240871021775544,
+      "grad_norm": 1.7265625,
+      "learning_rate": 3.0226314109032984e-07,
+      "loss": 0.020348772406578064,
+      "mean_token_accuracy": 0.9924048013985157,
+      "num_tokens": 45869501.0,
+      "step": 8620
+    },
+    {
+      "entropy": 0.01988984109498233,
+      "epoch": 0.9262311557788945,
+      "grad_norm": 2.09375,
+      "learning_rate": 2.8555445240461834e-07,
+      "loss": 0.01935010850429535,
+      "mean_token_accuracy": 0.9919997557997704,
+      "num_tokens": 45970153.0,
+      "step": 8640
+    },
+    {
+      "entropy": 0.01982094865072668,
+      "epoch": 0.9283752093802345,
+      "grad_norm": 2.09375,
+      "learning_rate": 2.6931411836595643e-07,
+      "loss": 0.02122329771518707,
+      "mean_token_accuracy": 0.9918153654783964,
+      "num_tokens": 46071225.0,
+      "step": 8660
+    },
+    {
+      "entropy": 0.019260422117486086,
+      "epoch": 0.9305192629815745,
+      "grad_norm": 2.359375,
+      "learning_rate": 2.535429219563401e-07,
+      "loss": 0.0204538568854332,
+      "mean_token_accuracy": 0.9913510870188474,
+      "num_tokens": 46180945.0,
+      "step": 8680
+    },
+    {
+      "entropy": 0.019332679143190035,
+      "epoch": 0.9326633165829146,
+      "grad_norm": 1.6796875,
+      "learning_rate": 2.3824162353960457e-07,
+      "loss": 0.02263512313365936,
+      "mean_token_accuracy": 0.9922845117747784,
+      "num_tokens": 46287275.0,
+      "step": 8700
+    },
+    {
+      "entropy": 0.018464213076777015,
+      "epoch": 0.9348073701842546,
+      "grad_norm": 2.046875,
+      "learning_rate": 2.2341096082478698e-07,
+      "loss": 0.019891925156116486,
+      "mean_token_accuracy": 0.9930620789527893,
+      "num_tokens": 46396075.0,
+      "step": 8720
+    },
+    {
+      "entropy": 0.020736305239142893,
+      "epoch": 0.9369514237855946,
+      "grad_norm": 1.796875,
+      "learning_rate": 2.0905164883054053e-07,
+      "loss": 0.0205291748046875,
+      "mean_token_accuracy": 0.9927501816302537,
+      "num_tokens": 46503373.0,
+      "step": 8740
+    },
+    {
+      "entropy": 0.01994141639102054,
+      "epoch": 0.9390954773869347,
+      "grad_norm": 1.8671875,
+      "learning_rate": 1.9516437985067306e-07,
+      "loss": 0.02209635227918625,
+      "mean_token_accuracy": 0.991724469140172,
+      "num_tokens": 46605784.0,
+      "step": 8760
+    },
+    {
+      "entropy": 0.018055033115933837,
+      "epoch": 0.9412395309882747,
+      "grad_norm": 1.4375,
+      "learning_rate": 1.8174982342076819e-07,
+      "loss": 0.02114318758249283,
+      "mean_token_accuracy": 0.9917370740324258,
+      "num_tokens": 46710311.0,
+      "step": 8780
+    },
+    {
+      "entropy": 0.017862834344521163,
+      "epoch": 0.9433835845896147,
+      "grad_norm": 1.40625,
+      "learning_rate": 1.6880862628590343e-07,
+      "loss": 0.017796410620212554,
+      "mean_token_accuracy": 0.9937189962714911,
+      "num_tokens": 46818472.0,
+      "step": 8800
+    },
+    {
+      "entropy": 0.01954027940284959,
+      "epoch": 0.9455276381909548,
+      "grad_norm": 1.6015625,
+      "learning_rate": 1.5634141236947174e-07,
+      "loss": 0.01990315467119217,
+      "mean_token_accuracy": 0.9926852386444807,
+      "num_tokens": 46912066.0,
+      "step": 8820
+    },
+    {
+      "entropy": 0.01788209332662518,
+      "epoch": 0.9476716917922948,
+      "grad_norm": 3.28125,
+      "learning_rate": 1.4434878274309782e-07,
+      "loss": 0.020390361547470093,
+      "mean_token_accuracy": 0.9922173459082841,
+      "num_tokens": 47012839.0,
+      "step": 8840
+    },
+    {
+      "entropy": 0.019225874226185623,
+      "epoch": 0.9498157453936349,
+      "grad_norm": 1.546875,
+      "learning_rate": 1.3283131559766016e-07,
+      "loss": 0.019317436218261718,
+      "mean_token_accuracy": 0.9922500483691692,
+      "num_tokens": 47118537.0,
+      "step": 8860
+    },
+    {
+      "entropy": 0.019644861581696205,
+      "epoch": 0.9519597989949748,
+      "grad_norm": 1.921875,
+      "learning_rate": 1.217895662154178e-07,
+      "loss": 0.02014264166355133,
+      "mean_token_accuracy": 0.9908678252249956,
+      "num_tokens": 47218518.0,
+      "step": 8880
+    },
+    {
+      "entropy": 0.019726947057893084,
+      "epoch": 0.9541038525963149,
+      "grad_norm": 1.859375,
+      "learning_rate": 1.1122406694323628e-07,
+      "loss": 0.021320286393165588,
+      "mean_token_accuracy": 0.9916484046727418,
+      "num_tokens": 47318647.0,
+      "step": 8900
+    },
+    {
+      "entropy": 0.019115359010970678,
+      "epoch": 0.956247906197655,
+      "grad_norm": 1.6875,
+      "learning_rate": 1.0113532716692131e-07,
+      "loss": 0.019532181322574615,
+      "mean_token_accuracy": 0.9924892302602529,
+      "num_tokens": 47423152.0,
+      "step": 8920
+    },
+    {
+      "entropy": 0.019615740292147167,
+      "epoch": 0.9583919597989949,
+      "grad_norm": 1.4921875,
+      "learning_rate": 9.152383328666415e-08,
+      "loss": 0.024791160225868226,
+      "mean_token_accuracy": 0.9896657336503267,
+      "num_tokens": 47525402.0,
+      "step": 8940
+    },
+    {
+      "entropy": 0.018387637558203097,
+      "epoch": 0.960536013400335,
+      "grad_norm": 1.5390625,
+      "learning_rate": 8.239004869358468e-08,
+      "loss": 0.018946020305156706,
+      "mean_token_accuracy": 0.9929318383336068,
+      "num_tokens": 47641223.0,
+      "step": 8960
+    },
+    {
+      "entropy": 0.018579465060270195,
+      "epoch": 0.9626800670016751,
+      "grad_norm": 2.625,
+      "learning_rate": 7.373441374739942e-08,
+      "loss": 0.021176393330097198,
+      "mean_token_accuracy": 0.9917469136416912,
+      "num_tokens": 47742034.0,
+      "step": 8980
+    },
+    {
+      "entropy": 0.019280895775682438,
+      "epoch": 0.964824120603015,
+      "grad_norm": 2.375,
+      "learning_rate": 6.555734575518058e-08,
+      "loss": 0.020877310633659364,
+      "mean_token_accuracy": 0.9919904433190823,
+      "num_tokens": 47836864.0,
+      "step": 9000
+    },
+    {
+      "epoch": 0.964824120603015,
+      "eval_entropy": 0.01948107976220975,
+      "eval_loss": 0.019369935616850853,
+      "eval_mean_token_accuracy": 0.9921530286471049,
+      "eval_num_tokens": 47836864.0,
+      "eval_runtime": 20.3121,
+      "eval_samples_per_second": 36.924,
+      "eval_steps_per_second": 18.462,
+      "step": 9000
+    }
+  ],
+  "logging_steps": 20,
+  "max_steps": 9329,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 1000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 5.864530685993165e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-9000/training_args.bin b/checkpoint-9000/training_args.bin
new file mode 100644
index 0000000..96cea35
--- /dev/null
+++ b/checkpoint-9000/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:896375d502906677c7c6ac4e94fcf73d56f682781d3d2addb3cca71640f162b7
+size 5777
diff --git a/checkpoint-9329/chat_template.jinja b/checkpoint-9329/chat_template.jinja
new file mode 100644
index 0000000..01be9b3
--- /dev/null
+++ b/checkpoint-9329/chat_template.jinja
@@ -0,0 +1,89 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if message.content is string %}
+        {%- set content = message.content %}
+    {%- else %}
+        {%- set content = '' %}
+    {%- endif %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in content %}
+                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is false %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/checkpoint-9329/config.json b/checkpoint-9329/config.json
new file mode 100644
index 0000000..1d219d6
--- /dev/null
+++ b/checkpoint-9329/config.json
@@ -0,0 +1,63 @@
+{
+  "architectures": [
+    "Qwen3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": null,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 40960,
+  "max_window_layers": 28,
+  "model_type": "qwen3",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 1000000,
+    "rope_type": "default"
+  },
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.7.0",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
diff --git a/checkpoint-9329/generation_config.json b/checkpoint-9329/generation_config.json
new file mode 100644
index 0000000..5ec133d
--- /dev/null
+++ b/checkpoint-9329/generation_config.json
@@ -0,0 +1,12 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "temperature": 0.6,
+  "top_k": 20,
+  "top_p": 0.95,
+  "transformers_version": "5.7.0"
+}
diff --git a/checkpoint-9329/model.safetensors b/checkpoint-9329/model.safetensors
new file mode 100644
index 0000000..6365eab
--- /dev/null
+++ b/checkpoint-9329/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32f3da47bd90115ac1bdfe0f5ab9ec22f7e234160c83bebd5a516fbf54466774
+size 3441185608
diff --git a/checkpoint-9329/optimizer.pt b/checkpoint-9329/optimizer.pt
new file mode 100644
index 0000000..1e1d5c7
--- /dev/null
+++ b/checkpoint-9329/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae5c10e1b138bc3ae15f33f0c9eee356b14320c80dbf55617d89b3c5fd001c94
+size 6882572207
diff --git a/checkpoint-9329/rng_state.pth b/checkpoint-9329/rng_state.pth
new file mode 100644
index 0000000..34de6a6
--- /dev/null
+++ b/checkpoint-9329/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed3753ab7977739b8eda494dd72defae5750f7283141b11a8f562160ba4c1a23
+size 14645
diff --git a/checkpoint-9329/scheduler.pt b/checkpoint-9329/scheduler.pt
new file mode 100644
index 0000000..e589eb4
--- /dev/null
+++ b/checkpoint-9329/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:816451833579af6813879ac219814814680365671d8b362c9aae623946b66bc8
+size 1465
diff --git a/checkpoint-9329/tokenizer.json b/checkpoint-9329/tokenizer.json
new file mode 100644
index 0000000..c7afbed
--- /dev/null
+++ b/checkpoint-9329/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
+size 11422650
diff --git a/checkpoint-9329/tokenizer_config.json b/checkpoint-9329/tokenizer_config.json
new file mode 100644
index 0000000..770e41d
--- /dev/null
+++ b/checkpoint-9329/tokenizer_config.json
@@ -0,0 +1,30 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "local_files_only": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}
diff --git a/checkpoint-9329/trainer_state.json b/checkpoint-9329/trainer_state.json
new file mode 100644
index 0000000..743b133
--- /dev/null
+++ b/checkpoint-9329/trainer_state.json
@@ -0,0 +1,4903 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 9329,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 0.3966331691481173,
+      "epoch": 0.0021440536013400337,
+      "grad_norm": 129.0,
+      "learning_rate": 1.3571428571428572e-06,
+      "loss": 3.2754512786865235,
+      "mean_token_accuracy": 0.7025479640811682,
+      "num_tokens": 100865.0,
+      "step": 20
+    },
+    {
+      "entropy": 0.39382178746163843,
+      "epoch": 0.004288107202680067,
+      "grad_norm": 72.5,
+      "learning_rate": 2.785714285714286e-06,
+      "loss": 2.5979455947875976,
+      "mean_token_accuracy": 0.7279976762831211,
+      "num_tokens": 209237.0,
+      "step": 40
+    },
+    {
+      "entropy": 0.49158020988106726,
+      "epoch": 0.006432160804020101,
+      "grad_norm": 31.0,
+      "learning_rate": 4.2142857142857145e-06,
+      "loss": 1.34407320022583,
+      "mean_token_accuracy": 0.7954290494322777,
+      "num_tokens": 322722.0,
+      "step": 60
+    },
+    {
+      "entropy": 0.37321625389158725,
+      "epoch": 0.008576214405360135,
+      "grad_norm": 9.6875,
+      "learning_rate": 5.6428571428571435e-06,
+      "loss": 0.308458948135376,
+      "mean_token_accuracy": 0.9197555683553219,
+      "num_tokens": 426557.0,
+      "step": 80
+    },
+    {
+      "entropy": 0.09664252077927812,
+      "epoch": 0.010720268006700168,
+      "grad_norm": 4.25,
+      "learning_rate": 7.0714285714285726e-06,
+      "loss": 0.08551757931709289,
+      "mean_token_accuracy": 0.9772517420351505,
+      "num_tokens": 538043.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.06678911932976916,
+      "epoch": 0.012864321608040201,
+      "grad_norm": 6.59375,
+      "learning_rate": 8.5e-06,
+      "loss": 0.05725666880607605,
+      "mean_token_accuracy": 0.978511106595397,
+      "num_tokens": 651143.0,
+      "step": 120
+    },
+    {
+      "entropy": 0.05399178955703974,
+      "epoch": 0.015008375209380235,
+      "grad_norm": 1.8515625,
+      "learning_rate": 9.92857142857143e-06,
+      "loss": 0.0476167619228363,
+      "mean_token_accuracy": 0.9818631950765848,
+      "num_tokens": 759324.0,
+      "step": 140
+    },
+    {
+      "entropy": 0.050143458176171406,
+      "epoch": 0.01715242881072027,
+      "grad_norm": 2.65625,
+      "learning_rate": 1.1357142857142858e-05,
+      "loss": 0.048712074756622314,
+      "mean_token_accuracy": 0.9795998100191354,
+      "num_tokens": 854771.0,
+      "step": 160
+    },
+    {
+      "entropy": 0.04800442266860046,
+      "epoch": 0.019296482412060303,
+      "grad_norm": 2.53125,
+      "learning_rate": 1.2785714285714286e-05,
+      "loss": 0.046506857872009276,
+      "mean_token_accuracy": 0.9819129385054112,
+      "num_tokens": 965263.0,
+      "step": 180
+    },
+    {
+      "entropy": 0.04474863488285337,
+      "epoch": 0.021440536013400336,
+      "grad_norm": 2.46875,
+      "learning_rate": 1.4214285714285716e-05,
+      "loss": 0.04230240285396576,
+      "mean_token_accuracy": 0.9835243009030818,
+      "num_tokens": 1067106.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.04506044688168913,
+      "epoch": 0.02358458961474037,
+      "grad_norm": 4.59375,
+      "learning_rate": 1.5642857142857143e-05,
+      "loss": 0.04433030188083649,
+      "mean_token_accuracy": 0.9820781596004963,
+      "num_tokens": 1181133.0,
+      "step": 220
+    },
+    {
+      "entropy": 0.04305547107360326,
+      "epoch": 0.025728643216080403,
+      "grad_norm": 1.7890625,
+      "learning_rate": 1.7071428571428573e-05,
+      "loss": 0.04099105894565582,
+      "mean_token_accuracy": 0.9831817526370287,
+      "num_tokens": 1301484.0,
+      "step": 240
+    },
+    {
+      "entropy": 0.04266677545965649,
+      "epoch": 0.027872696817420436,
+      "grad_norm": 2.34375,
+      "learning_rate": 1.8500000000000002e-05,
+      "loss": 0.04160386621952057,
+      "mean_token_accuracy": 0.9813323471695184,
+      "num_tokens": 1403952.0,
+      "step": 260
+    },
+    {
+      "entropy": 0.04101881538517773,
+      "epoch": 0.03001675041876047,
+      "grad_norm": 2.59375,
+      "learning_rate": 1.992857142857143e-05,
+      "loss": 0.042915910482406616,
+      "mean_token_accuracy": 0.9821092572063208,
+      "num_tokens": 1517588.0,
+      "step": 280
+    },
+    {
+      "entropy": 0.04049238133156905,
+      "epoch": 0.032160804020100506,
+      "grad_norm": 2.703125,
+      "learning_rate": 1.9999782442429437e-05,
+      "loss": 0.04166347980499267,
+      "mean_token_accuracy": 0.9828211534768343,
+      "num_tokens": 1629657.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.04230240093893371,
+      "epoch": 0.03430485762144054,
+      "grad_norm": 1.2421875,
+      "learning_rate": 1.9999083376151357e-05,
+      "loss": 0.041567260026931764,
+      "mean_token_accuracy": 0.9822345357388258,
+      "num_tokens": 1721145.0,
+      "step": 320
+    },
+    {
+      "entropy": 0.03799658405041555,
+      "epoch": 0.03644891122278057,
+      "grad_norm": 3.3125,
+      "learning_rate": 1.9997902232214542e-05,
+      "loss": 0.0413459450006485,
+      "mean_token_accuracy": 0.9830254316329956,
+      "num_tokens": 1820492.0,
+      "step": 340
+    },
+    {
+      "entropy": 0.041703186882659796,
+      "epoch": 0.038592964824120606,
+      "grad_norm": 2.65625,
+      "learning_rate": 1.999623906756452e-05,
+      "loss": 0.040859296917915344,
+      "mean_token_accuracy": 0.976021521165967,
+      "num_tokens": 1931348.0,
+      "step": 360
+    },
+    {
+      "entropy": 0.04301113227265887,
+      "epoch": 0.04073701842546064,
+      "grad_norm": 1.7734375,
+      "learning_rate": 1.9994093962386102e-05,
+      "loss": 0.04014136493206024,
+      "mean_token_accuracy": 0.9837368253618479,
+      "num_tokens": 2047917.0,
+      "step": 380
+    },
+    {
+      "entropy": 0.041214162003598176,
+      "epoch": 0.04288107202680067,
+      "grad_norm": 1.5078125,
+      "learning_rate": 1.9991467020099492e-05,
+      "loss": 0.04444628953933716,
+      "mean_token_accuracy": 0.9809734560549259,
+      "num_tokens": 2149033.0,
+      "step": 400
+    },
+    {
+      "entropy": 0.04038102811900899,
+      "epoch": 0.045025125628140705,
+      "grad_norm": 1.6796875,
+      "learning_rate": 1.998835836735531e-05,
+      "loss": 0.0405484139919281,
+      "mean_token_accuracy": 0.982579792290926,
+      "num_tokens": 2248970.0,
+      "step": 420
+    },
+    {
+      "entropy": 0.039991069887764755,
+      "epoch": 0.04716917922948074,
+      "grad_norm": 1.09375,
+      "learning_rate": 1.9984768154028508e-05,
+      "loss": 0.03874191641807556,
+      "mean_token_accuracy": 0.9835739009082317,
+      "num_tokens": 2350681.0,
+      "step": 440
+    },
+    {
+      "entropy": 0.03837411678614444,
+      "epoch": 0.04931323283082077,
+      "grad_norm": 1.4296875,
+      "learning_rate": 1.9980696553211108e-05,
+      "loss": 0.04068012237548828,
+      "mean_token_accuracy": 0.9836468525230885,
+      "num_tokens": 2455656.0,
+      "step": 460
+    },
+    {
+      "entropy": 0.03925063398201019,
+      "epoch": 0.051457286432160805,
+      "grad_norm": 2.140625,
+      "learning_rate": 1.9976143761203893e-05,
+      "loss": 0.04098735451698303,
+      "mean_token_accuracy": 0.9809495396912098,
+      "num_tokens": 2559532.0,
+      "step": 480
+    },
+    {
+      "entropy": 0.03997599564900156,
+      "epoch": 0.05360134003350084,
+      "grad_norm": 1.546875,
+      "learning_rate": 1.99711099975069e-05,
+      "loss": 0.03899767398834229,
+      "mean_token_accuracy": 0.9831298362463713,
+      "num_tokens": 2654243.0,
+      "step": 500
+    },
+    {
+      "epoch": 0.05360134003350084,
+      "eval_entropy": 0.038057172449305655,
+      "eval_loss": 0.03832634538412094,
+      "eval_mean_token_accuracy": 0.9846738157272339,
+      "eval_num_tokens": 2654243.0,
+      "eval_runtime": 20.3077,
+      "eval_samples_per_second": 36.932,
+      "eval_steps_per_second": 18.466,
+      "step": 500
+    },
+    {
+      "entropy": 0.040187044665799476,
+      "epoch": 0.05574539363484087,
+      "grad_norm": 3.796875,
+      "learning_rate": 1.9965595504808886e-05,
+      "loss": 0.03858818709850311,
+      "mean_token_accuracy": 0.9832367639988661,
+      "num_tokens": 2751255.0,
+      "step": 520
+    },
+    {
+      "entropy": 0.035708320523553994,
+      "epoch": 0.057889447236180905,
+      "grad_norm": 1.8515625,
+      "learning_rate": 1.9959600548975593e-05,
+      "loss": 0.035676625370979306,
+      "mean_token_accuracy": 0.9861446134746075,
+      "num_tokens": 2849554.0,
+      "step": 540
+    },
+    {
+      "entropy": 0.036129794413864144,
+      "epoch": 0.06003350083752094,
+      "grad_norm": 1.3828125,
+      "learning_rate": 1.9953125419036935e-05,
+      "loss": 0.03737228214740753,
+      "mean_token_accuracy": 0.9838687002658844,
+      "num_tokens": 2948412.0,
+      "step": 560
+    },
+    {
+      "entropy": 0.036765301373088734,
+      "epoch": 0.06217755443886097,
+      "grad_norm": 1.6640625,
+      "learning_rate": 1.994617042717308e-05,
+      "loss": 0.03550122678279877,
+      "mean_token_accuracy": 0.9847423356026411,
+      "num_tokens": 3048326.0,
+      "step": 580
+    },
+    {
+      "entropy": 0.03411570826137904,
+      "epoch": 0.06432160804020101,
+      "grad_norm": 2.40625,
+      "learning_rate": 1.9938735908699384e-05,
+      "loss": 0.036294499039649965,
+      "mean_token_accuracy": 0.9855027917772532,
+      "num_tokens": 3161774.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.0356013277938473,
+      "epoch": 0.06646566164154104,
+      "grad_norm": 1.84375,
+      "learning_rate": 1.993082222205023e-05,
+      "loss": 0.034555235505104066,
+      "mean_token_accuracy": 0.9851896297186613,
+      "num_tokens": 3273915.0,
+      "step": 620
+    },
+    {
+      "entropy": 0.0345983263629023,
+      "epoch": 0.06860971524288108,
+      "grad_norm": 2.34375,
+      "learning_rate": 1.9922429748761733e-05,
+      "loss": 0.03647570013999939,
+      "mean_token_accuracy": 0.9840687446296215,
+      "num_tokens": 3381550.0,
+      "step": 640
+    },
+    {
+      "entropy": 0.03738872102476307,
+      "epoch": 0.0707537688442211,
+      "grad_norm": 1.7265625,
+      "learning_rate": 1.9913558893453382e-05,
+      "loss": 0.036225003004074094,
+      "mean_token_accuracy": 0.9841504596173763,
+      "num_tokens": 3484439.0,
+      "step": 660
+    },
+    {
+      "entropy": 0.037293507948561455,
+      "epoch": 0.07289782244556114,
+      "grad_norm": 2.0625,
+      "learning_rate": 1.9904210083808485e-05,
+      "loss": 0.03906684815883636,
+      "mean_token_accuracy": 0.9821466624736785,
+      "num_tokens": 3588594.0,
+      "step": 680
+    },
+    {
+      "entropy": 0.03572316213103477,
+      "epoch": 0.07504187604690117,
+      "grad_norm": 1.3671875,
+      "learning_rate": 1.9894383770553594e-05,
+      "loss": 0.03395065367221832,
+      "mean_token_accuracy": 0.9846713658422231,
+      "num_tokens": 3690877.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.0373659498654888,
+      "epoch": 0.07718592964824121,
+      "grad_norm": 1.1328125,
+      "learning_rate": 1.988408042743673e-05,
+      "loss": 0.03977937400341034,
+      "mean_token_accuracy": 0.982515212148428,
+      "num_tokens": 3802248.0,
+      "step": 720
+    },
+    {
+      "entropy": 0.03811679463542532,
+      "epoch": 0.07932998324958124,
+      "grad_norm": 3.171875,
+      "learning_rate": 1.9873300551204588e-05,
+      "loss": 0.03767184019088745,
+      "mean_token_accuracy": 0.9842769052833319,
+      "num_tokens": 3904946.0,
+      "step": 740
+    },
+    {
+      "entropy": 0.03399244465199445,
+      "epoch": 0.08147403685092128,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.9862044661578554e-05,
+      "loss": 0.033424127101898196,
+      "mean_token_accuracy": 0.9856430813670158,
+      "num_tokens": 4011765.0,
+      "step": 760
+    },
+    {
+      "entropy": 0.03507241023471579,
+      "epoch": 0.0836180904522613,
+      "grad_norm": 2.328125,
+      "learning_rate": 1.985031330122966e-05,
+      "loss": 0.034555363655090335,
+      "mean_token_accuracy": 0.9848727196455002,
+      "num_tokens": 4119751.0,
+      "step": 780
+    },
+    {
+      "entropy": 0.03295613579248311,
+      "epoch": 0.08576214405360134,
+      "grad_norm": 2.015625,
+      "learning_rate": 1.9838107035752422e-05,
+      "loss": 0.03498654961585999,
+      "mean_token_accuracy": 0.9861093372106552,
+      "num_tokens": 4234054.0,
+      "step": 800
+    },
+    {
+      "entropy": 0.03774066482146736,
+      "epoch": 0.08790619765494137,
+      "grad_norm": 1.453125,
+      "learning_rate": 1.9825426453637578e-05,
+      "loss": 0.03699869811534882,
+      "mean_token_accuracy": 0.9836779370903969,
+      "num_tokens": 4345322.0,
+      "step": 820
+    },
+    {
+      "entropy": 0.03479030575981597,
+      "epoch": 0.09005025125628141,
+      "grad_norm": 1.8984375,
+      "learning_rate": 1.9812272166243694e-05,
+      "loss": 0.03539972007274628,
+      "mean_token_accuracy": 0.98408287987113,
+      "num_tokens": 4439690.0,
+      "step": 840
+    },
+    {
+      "entropy": 0.03287471420262591,
+      "epoch": 0.09219430485762144,
+      "grad_norm": 1.9375,
+      "learning_rate": 1.979864480776771e-05,
+      "loss": 0.033072105050086974,
+      "mean_token_accuracy": 0.9868043974041939,
+      "num_tokens": 4547696.0,
+      "step": 860
+    },
+    {
+      "entropy": 0.03525375899625942,
+      "epoch": 0.09433835845896148,
+      "grad_norm": 1.3984375,
+      "learning_rate": 1.9784545035214358e-05,
+      "loss": 0.034766626358032224,
+      "mean_token_accuracy": 0.985904373601079,
+      "num_tokens": 4656576.0,
+      "step": 880
+    },
+    {
+      "entropy": 0.03227176534710452,
+      "epoch": 0.0964824120603015,
+      "grad_norm": 1.46875,
+      "learning_rate": 1.976997352836448e-05,
+      "loss": 0.032944831252098086,
+      "mean_token_accuracy": 0.9853209290653467,
+      "num_tokens": 4763714.0,
+      "step": 900
+    },
+    {
+      "entropy": 0.03179633504041703,
+      "epoch": 0.09862646566164154,
+      "grad_norm": 1.5,
+      "learning_rate": 1.975493098974226e-05,
+      "loss": 0.033859679102897645,
+      "mean_token_accuracy": 0.9853155620396137,
+      "num_tokens": 4862673.0,
+      "step": 920
+    },
+    {
+      "entropy": 0.0342723391513573,
+      "epoch": 0.10077051926298157,
+      "grad_norm": 1.703125,
+      "learning_rate": 1.9739418144581357e-05,
+      "loss": 0.035592520236968996,
+      "mean_token_accuracy": 0.984546373412013,
+      "num_tokens": 4972462.0,
+      "step": 940
+    },
+    {
+      "entropy": 0.031624948812532236,
+      "epoch": 0.10291457286432161,
+      "grad_norm": 1.875,
+      "learning_rate": 1.9723435740789937e-05,
+      "loss": 0.03236276209354401,
+      "mean_token_accuracy": 0.9860632043331862,
+      "num_tokens": 5076752.0,
+      "step": 960
+    },
+    {
+      "entropy": 0.03294982312654611,
+      "epoch": 0.10505862646566164,
+      "grad_norm": 1.9140625,
+      "learning_rate": 1.97069845489146e-05,
+      "loss": 0.03309372365474701,
+      "mean_token_accuracy": 0.9859170567244291,
+      "num_tokens": 5194841.0,
+      "step": 980
+    },
+    {
+      "entropy": 0.03638018769561313,
+      "epoch": 0.10720268006700168,
+      "grad_norm": 1.3671875,
+      "learning_rate": 1.969006536210327e-05,
+      "loss": 0.036982369422912595,
+      "mean_token_accuracy": 0.9845436029136181,
+      "num_tokens": 5295536.0,
+      "step": 1000
+    },
+    {
+      "epoch": 0.10720268006700168,
+      "eval_entropy": 0.03662689963728189,
+      "eval_loss": 0.033028148114681244,
+      "eval_mean_token_accuracy": 0.9869652398427328,
+      "eval_num_tokens": 5295536.0,
+      "eval_runtime": 20.3094,
+      "eval_samples_per_second": 36.929,
+      "eval_steps_per_second": 18.464,
+      "step": 1000
+    },
+    {
+      "entropy": 0.035084289042788444,
+      "epoch": 0.1093467336683417,
+      "grad_norm": 1.4609375,
+      "learning_rate": 1.96726789960669e-05,
+      "loss": 0.03461971580982208,
+      "mean_token_accuracy": 0.9852123014628887,
+      "num_tokens": 5398244.0,
+      "step": 1020
+    },
+    {
+      "entropy": 0.03179455297176901,
+      "epoch": 0.11149078726968174,
+      "grad_norm": 1.3515625,
+      "learning_rate": 1.965482628904019e-05,
+      "loss": 0.03005821704864502,
+      "mean_token_accuracy": 0.9795859795063734,
+      "num_tokens": 5513313.0,
+      "step": 1040
+    },
+    {
+      "entropy": 0.028698108330354445,
+      "epoch": 0.11363484087102177,
+      "grad_norm": 3.4375,
+      "learning_rate": 1.963650810174116e-05,
+      "loss": 0.03329857885837555,
+      "mean_token_accuracy": 0.9862619318068028,
+      "num_tokens": 5612778.0,
+      "step": 1060
+    },
+    {
+      "entropy": 0.03370735585340299,
+      "epoch": 0.11577889447236181,
+      "grad_norm": 1.84375,
+      "learning_rate": 1.9617725317329644e-05,
+      "loss": 0.03324403166770935,
+      "mean_token_accuracy": 0.985461600497365,
+      "num_tokens": 5719033.0,
+      "step": 1080
+    },
+    {
+      "entropy": 0.035593700742174404,
+      "epoch": 0.11792294807370184,
+      "grad_norm": 1.8671875,
+      "learning_rate": 1.9598478841364727e-05,
+      "loss": 0.034142997860908506,
+      "mean_token_accuracy": 0.985479149222374,
+      "num_tokens": 5833790.0,
+      "step": 1100
+    },
+    {
+      "entropy": 0.029760706042361564,
+      "epoch": 0.12006700167504188,
+      "grad_norm": 1.234375,
+      "learning_rate": 1.957876960176106e-05,
+      "loss": 0.03303615748882294,
+      "mean_token_accuracy": 0.9864209692925214,
+      "num_tokens": 5942042.0,
+      "step": 1120
+    },
+    {
+      "entropy": 0.03119294007847202,
+      "epoch": 0.12221105527638192,
+      "grad_norm": 2.234375,
+      "learning_rate": 1.955859854874416e-05,
+      "loss": 0.03496485948562622,
+      "mean_token_accuracy": 0.9869707945734263,
+      "num_tokens": 6052169.0,
+      "step": 1140
+    },
+    {
+      "entropy": 0.032620664288697296,
+      "epoch": 0.12435510887772194,
+      "grad_norm": 2.1875,
+      "learning_rate": 1.9537966654804568e-05,
+      "loss": 0.032474163174629214,
+      "mean_token_accuracy": 0.985412934049964,
+      "num_tokens": 6160024.0,
+      "step": 1160
+    },
+    {
+      "entropy": 0.029878590396401707,
+      "epoch": 0.12649916247906198,
+      "grad_norm": 1.3203125,
+      "learning_rate": 1.9516874914650972e-05,
+      "loss": 0.028646692633628845,
+      "mean_token_accuracy": 0.9869820442050695,
+      "num_tokens": 6264476.0,
+      "step": 1180
+    },
+    {
+      "entropy": 0.03366943938890472,
+      "epoch": 0.12864321608040202,
+      "grad_norm": 2.390625,
+      "learning_rate": 1.9495324345162255e-05,
+      "loss": 0.03512271046638489,
+      "mean_token_accuracy": 0.9845213159918785,
+      "num_tokens": 6377469.0,
+      "step": 1200
+    },
+    {
+      "entropy": 0.03266510410194314,
+      "epoch": 0.13078726968174204,
+      "grad_norm": 1.625,
+      "learning_rate": 1.947331598533846e-05,
+      "loss": 0.03071320950984955,
+      "mean_token_accuracy": 0.9855085987597704,
+      "num_tokens": 6489484.0,
+      "step": 1220
+    },
+    {
+      "entropy": 0.02746599452148075,
+      "epoch": 0.13293132328308208,
+      "grad_norm": 2.03125,
+      "learning_rate": 1.9450850896250705e-05,
+      "loss": 0.029182371497154237,
+      "mean_token_accuracy": 0.9887533523142338,
+      "num_tokens": 6594306.0,
+      "step": 1240
+    },
+    {
+      "entropy": 0.029792005700437586,
+      "epoch": 0.13507537688442212,
+      "grad_norm": 1.8359375,
+      "learning_rate": 1.9427930160990026e-05,
+      "loss": 0.03335131406784057,
+      "mean_token_accuracy": 0.9871865373104811,
+      "num_tokens": 6701348.0,
+      "step": 1260
+    },
+    {
+      "entropy": 0.03408827395323897,
+      "epoch": 0.13721943048576216,
+      "grad_norm": 1.40625,
+      "learning_rate": 1.9404554884615153e-05,
+      "loss": 0.033723458647727966,
+      "mean_token_accuracy": 0.9851170387119055,
+      "num_tokens": 6812895.0,
+      "step": 1280
+    },
+    {
+      "entropy": 0.03265384387705126,
+      "epoch": 0.13936348408710217,
+      "grad_norm": 2.125,
+      "learning_rate": 1.9380726194099236e-05,
+      "loss": 0.032703128457069394,
+      "mean_token_accuracy": 0.9860071249306201,
+      "num_tokens": 6914997.0,
+      "step": 1300
+    },
+    {
+      "entropy": 0.030018781982835207,
+      "epoch": 0.1415075376884422,
+      "grad_norm": 1.9765625,
+      "learning_rate": 1.9356445238275518e-05,
+      "loss": 0.029086670279502867,
+      "mean_token_accuracy": 0.9875706855207682,
+      "num_tokens": 7017334.0,
+      "step": 1320
+    },
+    {
+      "entropy": 0.030701546505588338,
+      "epoch": 0.14365159128978225,
+      "grad_norm": 2.4375,
+      "learning_rate": 1.933171318778193e-05,
+      "loss": 0.03185203075408936,
+      "mean_token_accuracy": 0.9859924230724573,
+      "num_tokens": 7124252.0,
+      "step": 1340
+    },
+    {
+      "entropy": 0.032219707980402745,
+      "epoch": 0.1457956448911223,
+      "grad_norm": 1.359375,
+      "learning_rate": 1.9306531235004674e-05,
+      "loss": 0.03223040103912354,
+      "mean_token_accuracy": 0.9870269387960434,
+      "num_tokens": 7234001.0,
+      "step": 1360
+    },
+    {
+      "entropy": 0.029969332017935814,
+      "epoch": 0.1479396984924623,
+      "grad_norm": 1.578125,
+      "learning_rate": 1.928090059402072e-05,
+      "loss": 0.030724048614501953,
+      "mean_token_accuracy": 0.9877043284475804,
+      "num_tokens": 7340453.0,
+      "step": 1380
+    },
+    {
+      "entropy": 0.031696403899695726,
+      "epoch": 0.15008375209380234,
+      "grad_norm": 1.25,
+      "learning_rate": 1.9254822500539268e-05,
+      "loss": 0.032300516963005066,
+      "mean_token_accuracy": 0.9876017738133669,
+      "num_tokens": 7453174.0,
+      "step": 1400
+    },
+    {
+      "entropy": 0.0315240634514339,
+      "epoch": 0.15222780569514238,
+      "grad_norm": 1.328125,
+      "learning_rate": 1.9228298211842193e-05,
+      "loss": 0.033043709397315976,
+      "mean_token_accuracy": 0.9866235084831715,
+      "num_tokens": 7555849.0,
+      "step": 1420
+    },
+    {
+      "entropy": 0.031768915360407844,
+      "epoch": 0.15437185929648242,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.9201329006723413e-05,
+      "loss": 0.027889394760131837,
+      "mean_token_accuracy": 0.9881493493914604,
+      "num_tokens": 7665192.0,
+      "step": 1440
+    },
+    {
+      "entropy": 0.02793313514721376,
+      "epoch": 0.15651591289782243,
+      "grad_norm": 1.6015625,
+      "learning_rate": 1.917391618542724e-05,
+      "loss": 0.034602677822113036,
+      "mean_token_accuracy": 0.9850024841725826,
+      "num_tokens": 7766528.0,
+      "step": 1460
+    },
+    {
+      "entropy": 0.033910460214247,
+      "epoch": 0.15865996649916247,
+      "grad_norm": 2.375,
+      "learning_rate": 1.914606106958568e-05,
+      "loss": 0.03377147018909454,
+      "mean_token_accuracy": 0.9864576529711485,
+      "num_tokens": 7873249.0,
+      "step": 1480
+    },
+    {
+      "entropy": 0.03134421622089576,
+      "epoch": 0.16080402010050251,
+      "grad_norm": 1.6328125,
+      "learning_rate": 1.911776500215474e-05,
+      "loss": 0.031429946422576904,
+      "mean_token_accuracy": 0.9866740386933088,
+      "num_tokens": 7966983.0,
+      "step": 1500
+    },
+    {
+      "epoch": 0.16080402010050251,
+      "eval_entropy": 0.031583791558475544,
+      "eval_loss": 0.029090270400047302,
+      "eval_mean_token_accuracy": 0.9873407395680746,
+      "eval_num_tokens": 7966983.0,
+      "eval_runtime": 20.2978,
+      "eval_samples_per_second": 36.95,
+      "eval_steps_per_second": 18.475,
+      "step": 1500
+    },
+    {
+      "entropy": 0.03011373901317711,
+      "epoch": 0.16294807370184256,
+      "grad_norm": 1.96875,
+      "learning_rate": 1.9089029347349664e-05,
+      "loss": 0.029482910037040712,
+      "mean_token_accuracy": 0.9875899031758308,
+      "num_tokens": 8076659.0,
+      "step": 1520
+    },
+    {
+      "entropy": 0.024959181708436518,
+      "epoch": 0.16509212730318257,
+      "grad_norm": 2.875,
+      "learning_rate": 1.9059855490579156e-05,
+      "loss": 0.02683677673339844,
+      "mean_token_accuracy": 0.989487374201417,
+      "num_tokens": 8175314.0,
+      "step": 1540
+    },
+    {
+      "entropy": 0.027774957335896033,
+      "epoch": 0.1672361809045226,
+      "grad_norm": 2.09375,
+      "learning_rate": 1.90302448383786e-05,
+      "loss": 0.03151543140411377,
+      "mean_token_accuracy": 0.9872768469154835,
+      "num_tokens": 8290905.0,
+      "step": 1560
+    },
+    {
+      "entropy": 0.03308429398821318,
+      "epoch": 0.16938023450586265,
+      "grad_norm": 1.203125,
+      "learning_rate": 1.9000198818342243e-05,
+      "loss": 0.03073907792568207,
+      "mean_token_accuracy": 0.9868967544287444,
+      "num_tokens": 8397808.0,
+      "step": 1580
+    },
+    {
+      "entropy": 0.030384358397623146,
+      "epoch": 0.1715242881072027,
+      "grad_norm": 1.9296875,
+      "learning_rate": 1.8969718879054365e-05,
+      "loss": 0.034263554215431216,
+      "mean_token_accuracy": 0.98636131323874,
+      "num_tokens": 8504583.0,
+      "step": 1600
+    },
+    {
+      "entropy": 0.02909080486315361,
+      "epoch": 0.1736683417085427,
+      "grad_norm": 2.078125,
+      "learning_rate": 1.893880649001944e-05,
+      "loss": 0.029698848724365234,
+      "mean_token_accuracy": 0.9875615444034338,
+      "num_tokens": 8608889.0,
+      "step": 1620
+    },
+    {
+      "entropy": 0.03127812161601469,
+      "epoch": 0.17581239530988274,
+      "grad_norm": 2.3125,
+      "learning_rate": 1.890746314159129e-05,
+      "loss": 0.031479662656784056,
+      "mean_token_accuracy": 0.9861587047576904,
+      "num_tokens": 8720168.0,
+      "step": 1640
+    },
+    {
+      "entropy": 0.028632436878615407,
+      "epoch": 0.17795644891122278,
+      "grad_norm": 2.96875,
+      "learning_rate": 1.8875690344901237e-05,
+      "loss": 0.02911217212677002,
+      "mean_token_accuracy": 0.9880903176963329,
+      "num_tokens": 8822057.0,
+      "step": 1660
+    },
+    {
+      "entropy": 0.02712122074663057,
+      "epoch": 0.18010050251256282,
+      "grad_norm": 5.09375,
+      "learning_rate": 1.884348963178523e-05,
+      "loss": 0.02785937786102295,
+      "mean_token_accuracy": 0.9886827927082777,
+      "num_tokens": 8926336.0,
+      "step": 1680
+    },
+    {
+      "entropy": 0.025899743908667006,
+      "epoch": 0.18224455611390286,
+      "grad_norm": 1.265625,
+      "learning_rate": 1.881086255471001e-05,
+      "loss": 0.026762637495994567,
+      "mean_token_accuracy": 0.9886533226817846,
+      "num_tokens": 9023646.0,
+      "step": 1700
+    },
+    {
+      "entropy": 0.028882844069084968,
+      "epoch": 0.18438860971524287,
+      "grad_norm": 1.5390625,
+      "learning_rate": 1.8777810686698266e-05,
+      "loss": 0.029647400975227355,
+      "mean_token_accuracy": 0.9875607810914516,
+      "num_tokens": 9127152.0,
+      "step": 1720
+    },
+    {
+      "entropy": 0.03044446731946664,
+      "epoch": 0.18653266331658291,
+      "grad_norm": 1.2265625,
+      "learning_rate": 1.8744335621252766e-05,
+      "loss": 0.031170135736465453,
+      "mean_token_accuracy": 0.9866547036916018,
+      "num_tokens": 9234279.0,
+      "step": 1740
+    },
+    {
+      "entropy": 0.029704924078851037,
+      "epoch": 0.18867671691792295,
+      "grad_norm": 1.546875,
+      "learning_rate": 1.8710438972279572e-05,
+      "loss": 0.030961766839027405,
+      "mean_token_accuracy": 0.9872025232762098,
+      "num_tokens": 9348546.0,
+      "step": 1760
+    },
+    {
+      "entropy": 0.026529241763273603,
+      "epoch": 0.190820770519263,
+      "grad_norm": 2.828125,
+      "learning_rate": 1.867612237401019e-05,
+      "loss": 0.026679041981697082,
+      "mean_token_accuracy": 0.9885351285338402,
+      "num_tokens": 9455489.0,
+      "step": 1780
+    },
+    {
+      "entropy": 0.028892852607532403,
+      "epoch": 0.192964824120603,
+      "grad_norm": 4.03125,
+      "learning_rate": 1.8641387480922816e-05,
+      "loss": 0.03060615360736847,
+      "mean_token_accuracy": 0.9880913089960813,
+      "num_tokens": 9561715.0,
+      "step": 1800
+    },
+    {
+      "entropy": 0.03127149782612833,
+      "epoch": 0.19510887772194305,
+      "grad_norm": 1.4375,
+      "learning_rate": 1.860623596766254e-05,
+      "loss": 0.03185727298259735,
+      "mean_token_accuracy": 0.9862949691712857,
+      "num_tokens": 9671395.0,
+      "step": 1820
+    },
+    {
+      "entropy": 0.030052698707368108,
+      "epoch": 0.1972529313232831,
+      "grad_norm": 1.7265625,
+      "learning_rate": 1.8570669528960627e-05,
+      "loss": 0.030486562848091127,
+      "mean_token_accuracy": 0.9874879900366068,
+      "num_tokens": 9772135.0,
+      "step": 1840
+    },
+    {
+      "entropy": 0.028759777540108188,
+      "epoch": 0.19939698492462313,
+      "grad_norm": 1.5546875,
+      "learning_rate": 1.8534689879552806e-05,
+      "loss": 0.02927035093307495,
+      "mean_token_accuracy": 0.9883296392858029,
+      "num_tokens": 9875332.0,
+      "step": 1860
+    },
+    {
+      "entropy": 0.026194313674932346,
+      "epoch": 0.20154103852596314,
+      "grad_norm": 1.2421875,
+      "learning_rate": 1.8498298754096597e-05,
+      "loss": 0.027384281158447266,
+      "mean_token_accuracy": 0.9899293314665556,
+      "num_tokens": 9975866.0,
+      "step": 1880
+    },
+    {
+      "entropy": 0.030918022169134928,
+      "epoch": 0.20368509212730318,
+      "grad_norm": 1.7421875,
+      "learning_rate": 1.8461497907087672e-05,
+      "loss": 0.03030203580856323,
+      "mean_token_accuracy": 0.9866726931184531,
+      "num_tokens": 10091897.0,
+      "step": 1900
+    },
+    {
+      "entropy": 0.02552133451026748,
+      "epoch": 0.20582914572864322,
+      "grad_norm": 1.1875,
+      "learning_rate": 1.8424289112775284e-05,
+      "loss": 0.027020731568336488,
+      "mean_token_accuracy": 0.9879696272313595,
+      "num_tokens": 10189590.0,
+      "step": 1920
+    },
+    {
+      "entropy": 0.026746707828442594,
+      "epoch": 0.20797319932998326,
+      "grad_norm": 3.59375,
+      "learning_rate": 1.8386674165076706e-05,
+      "loss": 0.026832282543182373,
+      "mean_token_accuracy": 0.9871037002652884,
+      "num_tokens": 10290612.0,
+      "step": 1940
+    },
+    {
+      "entropy": 0.023915313509087356,
+      "epoch": 0.21011725293132327,
+      "grad_norm": 3.75,
+      "learning_rate": 1.8348654877490765e-05,
+      "loss": 0.03067324459552765,
+      "mean_token_accuracy": 0.9885836262255907,
+      "num_tokens": 10400962.0,
+      "step": 1960
+    },
+    {
+      "entropy": 0.027086517597672354,
+      "epoch": 0.2122613065326633,
+      "grad_norm": 2.078125,
+      "learning_rate": 1.831023308301039e-05,
+      "loss": 0.028256928920745848,
+      "mean_token_accuracy": 0.9885173581540585,
+      "num_tokens": 10506362.0,
+      "step": 1980
+    },
+    {
+      "entropy": 0.02741672217816813,
+      "epoch": 0.21440536013400335,
+      "grad_norm": 0.9140625,
+      "learning_rate": 1.827141063403425e-05,
+      "loss": 0.02612453103065491,
+      "mean_token_accuracy": 0.9899405423551798,
+      "num_tokens": 10614125.0,
+      "step": 2000
+    },
+    {
+      "epoch": 0.21440536013400335,
+      "eval_entropy": 0.024848208941170014,
+      "eval_loss": 0.026614099740982056,
+      "eval_mean_token_accuracy": 0.9895401794115702,
+      "eval_num_tokens": 10614125.0,
+      "eval_runtime": 20.3696,
+      "eval_samples_per_second": 36.82,
+      "eval_steps_per_second": 18.41,
+      "step": 2000
+    },
+    {
+      "entropy": 0.024184393659743363,
+      "epoch": 0.2165494137353434,
+      "grad_norm": 1.1015625,
+      "learning_rate": 1.8232189402277446e-05,
+      "loss": 0.029155582189559937,
+      "mean_token_accuracy": 0.9877722147852183,
+      "num_tokens": 10718545.0,
+      "step": 2020
+    },
+    {
+      "entropy": 0.026252732726061368,
+      "epoch": 0.2186934673366834,
+      "grad_norm": 1.8046875,
+      "learning_rate": 1.8192571278681253e-05,
+      "loss": 0.026098987460136412,
+      "mean_token_accuracy": 0.9895182501524686,
+      "num_tokens": 10826477.0,
+      "step": 2040
+    },
+    {
+      "entropy": 0.025545999853238754,
+      "epoch": 0.22083752093802345,
+      "grad_norm": 2.234375,
+      "learning_rate": 1.8152558173321987e-05,
+      "loss": 0.026094555854797363,
+      "mean_token_accuracy": 0.9890841003507376,
+      "num_tokens": 10925874.0,
+      "step": 2060
+    },
+    {
+      "entropy": 0.026384522972375635,
+      "epoch": 0.2229815745393635,
+      "grad_norm": 2.40625,
+      "learning_rate": 1.8112152015318894e-05,
+      "loss": 0.029251503944396972,
+      "mean_token_accuracy": 0.9880545783787966,
+      "num_tokens": 11033510.0,
+      "step": 2080
+    },
+    {
+      "entropy": 0.025210863275060547,
+      "epoch": 0.22512562814070353,
+      "grad_norm": 2.703125,
+      "learning_rate": 1.8071354752741137e-05,
+      "loss": 0.027124592661857606,
+      "mean_token_accuracy": 0.9892004739493132,
+      "num_tokens": 11137817.0,
+      "step": 2100
+    },
+    {
+      "entropy": 0.027089714014164202,
+      "epoch": 0.22726968174204354,
+      "grad_norm": 1.7578125,
+      "learning_rate": 1.8030168352513894e-05,
+      "loss": 0.026511949300765992,
+      "mean_token_accuracy": 0.9889535740017891,
+      "num_tokens": 11245185.0,
+      "step": 2120
+    },
+    {
+      "entropy": 0.024535768478563112,
+      "epoch": 0.22941373534338358,
+      "grad_norm": 4.625,
+      "learning_rate": 1.7988594800323515e-05,
+      "loss": 0.02658265233039856,
+      "mean_token_accuracy": 0.9905748564749957,
+      "num_tokens": 11349330.0,
+      "step": 2140
+    },
+    {
+      "entropy": 0.02754759546187415,
+      "epoch": 0.23155778894472362,
+      "grad_norm": 1.9609375,
+      "learning_rate": 1.7946636100521798e-05,
+      "loss": 0.027984172105789185,
+      "mean_token_accuracy": 0.9884784188121557,
+      "num_tokens": 11461079.0,
+      "step": 2160
+    },
+    {
+      "entropy": 0.025805097624288464,
+      "epoch": 0.23370184254606366,
+      "grad_norm": 2.078125,
+      "learning_rate": 1.7904294276029346e-05,
+      "loss": 0.026595547795295715,
+      "mean_token_accuracy": 0.9894516475498676,
+      "num_tokens": 11563866.0,
+      "step": 2180
+    },
+    {
+      "entropy": 0.026482580404353938,
+      "epoch": 0.23584589614740367,
+      "grad_norm": 2.765625,
+      "learning_rate": 1.7861571368238035e-05,
+      "loss": 0.02969781160354614,
+      "mean_token_accuracy": 0.9875592797994613,
+      "num_tokens": 11669471.0,
+      "step": 2200
+    },
+    {
+      "entropy": 0.028677671598052255,
+      "epoch": 0.2379899497487437,
+      "grad_norm": 2.21875,
+      "learning_rate": 1.7818469436912605e-05,
+      "loss": 0.02763218283653259,
+      "mean_token_accuracy": 0.9888417005538941,
+      "num_tokens": 11768609.0,
+      "step": 2220
+    },
+    {
+      "entropy": 0.026651295177725843,
+      "epoch": 0.24013400335008375,
+      "grad_norm": 1.859375,
+      "learning_rate": 1.7774990560091356e-05,
+      "loss": 0.02962457537651062,
+      "mean_token_accuracy": 0.9879184097051621,
+      "num_tokens": 11880854.0,
+      "step": 2240
+    },
+    {
+      "entropy": 0.02831397352310887,
+      "epoch": 0.2422780569514238,
+      "grad_norm": 1.8984375,
+      "learning_rate": 1.773113683398593e-05,
+      "loss": 0.0296058714389801,
+      "mean_token_accuracy": 0.9883440136909485,
+      "num_tokens": 11986546.0,
+      "step": 2260
+    },
+    {
+      "entropy": 0.0280173282616488,
+      "epoch": 0.24442211055276383,
+      "grad_norm": 1.9609375,
+      "learning_rate": 1.76869103728803e-05,
+      "loss": 0.026104235649108888,
+      "mean_token_accuracy": 0.9899998947978019,
+      "num_tokens": 12089099.0,
+      "step": 2280
+    },
+    {
+      "entropy": 0.027054303306613292,
+      "epoch": 0.24656616415410385,
+      "grad_norm": 1.7734375,
+      "learning_rate": 1.7642313309028798e-05,
+      "loss": 0.029236692190170287,
+      "mean_token_accuracy": 0.988573432713747,
+      "num_tokens": 12198358.0,
+      "step": 2300
+    },
+    {
+      "entropy": 0.025954095610904913,
+      "epoch": 0.24871021775544389,
+      "grad_norm": 2.53125,
+      "learning_rate": 1.7597347792553313e-05,
+      "loss": 0.02622365057468414,
+      "mean_token_accuracy": 0.9882505975663662,
+      "num_tokens": 12310140.0,
+      "step": 2320
+    },
+    {
+      "entropy": 0.02558089109697903,
+      "epoch": 0.2508542713567839,
+      "grad_norm": 2.5625,
+      "learning_rate": 1.7552015991339655e-05,
+      "loss": 0.025320738554000854,
+      "mean_token_accuracy": 0.9904545627534389,
+      "num_tokens": 12413395.0,
+      "step": 2340
+    },
+    {
+      "entropy": 0.02581585149782768,
+      "epoch": 0.25299832495812397,
+      "grad_norm": 1.59375,
+      "learning_rate": 1.7506320090933012e-05,
+      "loss": 0.02813691198825836,
+      "mean_token_accuracy": 0.9890731148421764,
+      "num_tokens": 12517491.0,
+      "step": 2360
+    },
+    {
+      "entropy": 0.029305921567720362,
+      "epoch": 0.255142378559464,
+      "grad_norm": 1.46875,
+      "learning_rate": 1.74602622944326e-05,
+      "loss": 0.030352601408958436,
+      "mean_token_accuracy": 0.9881349734961986,
+      "num_tokens": 12627025.0,
+      "step": 2380
+    },
+    {
+      "entropy": 0.026154012439656073,
+      "epoch": 0.25728643216080405,
+      "grad_norm": 2.03125,
+      "learning_rate": 1.7413844822385423e-05,
+      "loss": 0.0258339524269104,
+      "mean_token_accuracy": 0.9894504573196172,
+      "num_tokens": 12726236.0,
+      "step": 2400
+    },
+    {
+      "entropy": 0.024479785768926378,
+      "epoch": 0.25943048576214406,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.736706991267924e-05,
+      "loss": 0.027092832326889037,
+      "mean_token_accuracy": 0.9882367230951786,
+      "num_tokens": 12834614.0,
+      "step": 2420
+    },
+    {
+      "entropy": 0.02744944194200798,
+      "epoch": 0.26157453936348407,
+      "grad_norm": 2.734375,
+      "learning_rate": 1.7319939820434656e-05,
+      "loss": 0.026963105797767638,
+      "mean_token_accuracy": 0.9900378599762917,
+      "num_tokens": 12940452.0,
+      "step": 2440
+    },
+    {
+      "entropy": 0.02669895636208821,
+      "epoch": 0.26371859296482414,
+      "grad_norm": 6.9375,
+      "learning_rate": 1.72724568178964e-05,
+      "loss": 0.02737392783164978,
+      "mean_token_accuracy": 0.9884641785174608,
+      "num_tokens": 13038623.0,
+      "step": 2460
+    },
+    {
+      "entropy": 0.02623358482396725,
+      "epoch": 0.26586264656616415,
+      "grad_norm": 1.65625,
+      "learning_rate": 1.7224623194323785e-05,
+      "loss": 0.026823097467422487,
+      "mean_token_accuracy": 0.9890745796263218,
+      "num_tokens": 13145771.0,
+      "step": 2480
+    },
+    {
+      "entropy": 0.02654599146626424,
+      "epoch": 0.26800670016750416,
+      "grad_norm": 2.125,
+      "learning_rate": 1.717644125588032e-05,
+      "loss": 0.02797144651412964,
+      "mean_token_accuracy": 0.9888463616371155,
+      "num_tokens": 13258191.0,
+      "step": 2500
+    },
+    {
+      "epoch": 0.26800670016750416,
+      "eval_entropy": 0.02707736179015289,
+      "eval_loss": 0.024532999843358994,
+      "eval_mean_token_accuracy": 0.9901215243339538,
+      "eval_num_tokens": 13258191.0,
+      "eval_runtime": 20.2919,
+      "eval_samples_per_second": 36.961,
+      "eval_steps_per_second": 18.48,
+      "step": 2500
+    },
+    {
+      "entropy": 0.027981107782488836,
+      "epoch": 0.27015075376884423,
+      "grad_norm": 2.140625,
+      "learning_rate": 1.7127913325522546e-05,
+      "loss": 0.027564391493797302,
+      "mean_token_accuracy": 0.9884749103337527,
+      "num_tokens": 13360032.0,
+      "step": 2520
+    },
+    {
+      "entropy": 0.024831563800216828,
+      "epoch": 0.27229480737018424,
+      "grad_norm": 2.625,
+      "learning_rate": 1.7079041742888023e-05,
+      "loss": 0.027046608924865722,
+      "mean_token_accuracy": 0.9889734648168087,
+      "num_tokens": 13459997.0,
+      "step": 2540
+    },
+    {
+      "entropy": 0.025509285347288822,
+      "epoch": 0.2744388609715243,
+      "grad_norm": 2.875,
+      "learning_rate": 1.702982886418254e-05,
+      "loss": 0.026221618056297302,
+      "mean_token_accuracy": 0.9900430839508771,
+      "num_tokens": 13562047.0,
+      "step": 2560
+    },
+    {
+      "entropy": 0.02474073741786924,
+      "epoch": 0.2765829145728643,
+      "grad_norm": 2.875,
+      "learning_rate": 1.6980277062066513e-05,
+      "loss": 0.02481588125228882,
+      "mean_token_accuracy": 0.990446874499321,
+      "num_tokens": 13659557.0,
+      "step": 2580
+    },
+    {
+      "entropy": 0.026416731743847777,
+      "epoch": 0.27872696817420434,
+      "grad_norm": 1.7734375,
+      "learning_rate": 1.69303887255406e-05,
+      "loss": 0.026889228820800783,
+      "mean_token_accuracy": 0.9888020649552345,
+      "num_tokens": 13773310.0,
+      "step": 2600
+    },
+    {
+      "entropy": 0.0239254101291408,
+      "epoch": 0.2808710217755444,
+      "grad_norm": 1.4375,
+      "learning_rate": 1.6880166259830516e-05,
+      "loss": 0.02351238876581192,
+      "mean_token_accuracy": 0.9907949134707451,
+      "num_tokens": 13893984.0,
+      "step": 2620
+    },
+    {
+      "entropy": 0.02596387622606926,
+      "epoch": 0.2830150753768844,
+      "grad_norm": 1.859375,
+      "learning_rate": 1.682961208627108e-05,
+      "loss": 0.025493481755256654,
+      "mean_token_accuracy": 0.9891014736145735,
+      "num_tokens": 13994782.0,
+      "step": 2640
+    },
+    {
+      "entropy": 0.023969465121626853,
+      "epoch": 0.28515912897822443,
+      "grad_norm": 4.21875,
+      "learning_rate": 1.6778728642189456e-05,
+      "loss": 0.025132390856742858,
+      "mean_token_accuracy": 0.9904550548642874,
+      "num_tokens": 14106793.0,
+      "step": 2660
+    },
+    {
+      "entropy": 0.023048270364415657,
+      "epoch": 0.2873031825795645,
+      "grad_norm": 3.390625,
+      "learning_rate": 1.6727518380787678e-05,
+      "loss": 0.026464682817459107,
+      "mean_token_accuracy": 0.98928731828928,
+      "num_tokens": 14215090.0,
+      "step": 2680
+    },
+    {
+      "entropy": 0.024989640445710393,
+      "epoch": 0.2894472361809045,
+      "grad_norm": 1.0546875,
+      "learning_rate": 1.6675983771024345e-05,
+      "loss": 0.02458348125219345,
+      "mean_token_accuracy": 0.9901158206164837,
+      "num_tokens": 14321065.0,
+      "step": 2700
+    },
+    {
+      "entropy": 0.023273660653694607,
+      "epoch": 0.2915912897822446,
+      "grad_norm": 2.203125,
+      "learning_rate": 1.6624127297495606e-05,
+      "loss": 0.02475869804620743,
+      "mean_token_accuracy": 0.9907850239425897,
+      "num_tokens": 14421911.0,
+      "step": 2720
+    },
+    {
+      "entropy": 0.02422142572395387,
+      "epoch": 0.2937353433835846,
+      "grad_norm": 2.015625,
+      "learning_rate": 1.6571951460315362e-05,
+      "loss": 0.027780982851982116,
+      "mean_token_accuracy": 0.9886230506002903,
+      "num_tokens": 14529639.0,
+      "step": 2740
+    },
+    {
+      "entropy": 0.0267424136016416,
+      "epoch": 0.2958793969849246,
+      "grad_norm": 1.9765625,
+      "learning_rate": 1.6519458774994728e-05,
+      "loss": 0.026475533843040466,
+      "mean_token_accuracy": 0.9882622588425874,
+      "num_tokens": 14634262.0,
+      "step": 2760
+    },
+    {
+      "entropy": 0.023857305470937718,
+      "epoch": 0.29802345058626467,
+      "grad_norm": 2.390625,
+      "learning_rate": 1.646665177232077e-05,
+      "loss": 0.026801908016204835,
+      "mean_token_accuracy": 0.9902123864740133,
+      "num_tokens": 14739645.0,
+      "step": 2780
+    },
+    {
+      "entropy": 0.027312484692993168,
+      "epoch": 0.3001675041876047,
+      "grad_norm": 1.9765625,
+      "learning_rate": 1.6413532998234477e-05,
+      "loss": 0.027185559272766113,
+      "mean_token_accuracy": 0.9890467882156372,
+      "num_tokens": 14844155.0,
+      "step": 2800
+    },
+    {
+      "entropy": 0.02615051964248778,
+      "epoch": 0.3023115577889447,
+      "grad_norm": 1.859375,
+      "learning_rate": 1.6360105013708022e-05,
+      "loss": 0.02619759738445282,
+      "mean_token_accuracy": 0.9885728180408477,
+      "num_tokens": 14947978.0,
+      "step": 2820
+    },
+    {
+      "entropy": 0.023987949729507817,
+      "epoch": 0.30445561139028476,
+      "grad_norm": 1.2734375,
+      "learning_rate": 1.6306370394621274e-05,
+      "loss": 0.02713790535926819,
+      "mean_token_accuracy": 0.989062811806798,
+      "num_tokens": 15043851.0,
+      "step": 2840
+    },
+    {
+      "entropy": 0.024998777841574338,
+      "epoch": 0.3065996649916248,
+      "grad_norm": 2.59375,
+      "learning_rate": 1.625233173163765e-05,
+      "loss": 0.02687191665172577,
+      "mean_token_accuracy": 0.989005958288908,
+      "num_tokens": 15168916.0,
+      "step": 2860
+    },
+    {
+      "entropy": 0.025794855099957204,
+      "epoch": 0.30874371859296484,
+      "grad_norm": 1.296875,
+      "learning_rate": 1.619799163007916e-05,
+      "loss": 0.02666613459587097,
+      "mean_token_accuracy": 0.989616809785366,
+      "num_tokens": 15283048.0,
+      "step": 2880
+    },
+    {
+      "entropy": 0.023574712648041896,
+      "epoch": 0.31088777219430486,
+      "grad_norm": 1.2578125,
+      "learning_rate": 1.6143352709800847e-05,
+      "loss": 0.022058571875095367,
+      "mean_token_accuracy": 0.9904920358210803,
+      "num_tokens": 15401868.0,
+      "step": 2900
+    },
+    {
+      "entropy": 0.02349452629246116,
+      "epoch": 0.31303182579564487,
+      "grad_norm": 2.375,
+      "learning_rate": 1.6088417605064453e-05,
+      "loss": 0.025521630048751832,
+      "mean_token_accuracy": 0.9904454626142979,
+      "num_tokens": 15512377.0,
+      "step": 2920
+    },
+    {
+      "entropy": 0.024990445605726563,
+      "epoch": 0.31517587939698494,
+      "grad_norm": 1.7734375,
+      "learning_rate": 1.6033188964411407e-05,
+      "loss": 0.027597138285636903,
+      "mean_token_accuracy": 0.9910409476608038,
+      "num_tokens": 15626267.0,
+      "step": 2940
+    },
+    {
+      "entropy": 0.022828791862048094,
+      "epoch": 0.31731993299832495,
+      "grad_norm": 3.65625,
+      "learning_rate": 1.5977669450535164e-05,
+      "loss": 0.023614174127578734,
+      "mean_token_accuracy": 0.9908586539328098,
+      "num_tokens": 15732131.0,
+      "step": 2960
+    },
+    {
+      "entropy": 0.025171676824174937,
+      "epoch": 0.319463986599665,
+      "grad_norm": 1.0546875,
+      "learning_rate": 1.5921861740152803e-05,
+      "loss": 0.029384702444076538,
+      "mean_token_accuracy": 0.9898778181523085,
+      "num_tokens": 15827873.0,
+      "step": 2980
+    },
+    {
+      "entropy": 0.025553616087654517,
+      "epoch": 0.32160804020100503,
+      "grad_norm": 1.8828125,
+      "learning_rate": 1.586576852387599e-05,
+      "loss": 0.022602304816246033,
+      "mean_token_accuracy": 0.9912034373730421,
+      "num_tokens": 15920514.0,
+      "step": 3000
+    },
+    {
+      "epoch": 0.32160804020100503,
+      "eval_entropy": 0.024538942879318106,
+      "eval_loss": 0.02338256873190403,
+      "eval_mean_token_accuracy": 0.9907407927513122,
+      "eval_num_tokens": 15920514.0,
+      "eval_runtime": 20.2778,
+      "eval_samples_per_second": 36.986,
+      "eval_steps_per_second": 18.493,
+      "step": 3000
+    },
+    {
+      "entropy": 0.021682905555485375,
+      "epoch": 0.32375209380234504,
+      "grad_norm": 2.46875,
+      "learning_rate": 1.5809392506081244e-05,
+      "loss": 0.022935517132282257,
+      "mean_token_accuracy": 0.9904364123940468,
+      "num_tokens": 16035334.0,
+      "step": 3020
+    },
+    {
+      "entropy": 0.01898658178452024,
+      "epoch": 0.3258961474036851,
+      "grad_norm": 1.6796875,
+      "learning_rate": 1.5752736404779577e-05,
+      "loss": 0.02092217206954956,
+      "mean_token_accuracy": 0.984833637624979,
+      "num_tokens": 16140327.0,
+      "step": 3040
+    },
+    {
+      "entropy": 0.021216248062864908,
+      "epoch": 0.3280402010050251,
+      "grad_norm": 2.3125,
+      "learning_rate": 1.5695802951485434e-05,
+      "loss": 0.024958278238773345,
+      "mean_token_accuracy": 0.9900662276893855,
+      "num_tokens": 16255099.0,
+      "step": 3060
+    },
+    {
+      "entropy": 0.023438269265625422,
+      "epoch": 0.33018425460636514,
+      "grad_norm": 1.7421875,
+      "learning_rate": 1.5638594891085e-05,
+      "loss": 0.024300065636634827,
+      "mean_token_accuracy": 0.9898340333253145,
+      "num_tokens": 16366533.0,
+      "step": 3080
+    },
+    {
+      "entropy": 0.02112419887921533,
+      "epoch": 0.3323283082077052,
+      "grad_norm": 1.5078125,
+      "learning_rate": 1.5581114981703878e-05,
+      "loss": 0.02245298624038696,
+      "mean_token_accuracy": 0.9914943426847458,
+      "num_tokens": 16469428.0,
+      "step": 3100
+    },
+    {
+      "entropy": 0.023446582856286112,
+      "epoch": 0.3344723618090452,
+      "grad_norm": 2.78125,
+      "learning_rate": 1.5523365994574098e-05,
+      "loss": 0.024820271134376525,
+      "mean_token_accuracy": 0.990750040858984,
+      "num_tokens": 16582498.0,
+      "step": 3120
+    },
+    {
+      "entropy": 0.022754177751903627,
+      "epoch": 0.3366164154103853,
+      "grad_norm": 1.8203125,
+      "learning_rate": 1.546535071390052e-05,
+      "loss": 0.025619438290596007,
+      "mean_token_accuracy": 0.9895062685012818,
+      "num_tokens": 16693045.0,
+      "step": 3140
+    },
+    {
+      "entropy": 0.023310942134048673,
+      "epoch": 0.3387604690117253,
+      "grad_norm": 2.546875,
+      "learning_rate": 1.540707193672659e-05,
+      "loss": 0.02326971888542175,
+      "mean_token_accuracy": 0.9908227741718292,
+      "num_tokens": 16796636.0,
+      "step": 3160
+    },
+    {
+      "entropy": 0.02383078648617811,
+      "epoch": 0.3409045226130653,
+      "grad_norm": 3.390625,
+      "learning_rate": 1.53485324727995e-05,
+      "loss": 0.025269970297813416,
+      "mean_token_accuracy": 0.9894092302769423,
+      "num_tokens": 16908823.0,
+      "step": 3180
+    },
+    {
+      "entropy": 0.025040690238802198,
+      "epoch": 0.3430485762144054,
+      "grad_norm": 2.515625,
+      "learning_rate": 1.5289735144434733e-05,
+      "loss": 0.025703960657119752,
+      "mean_token_accuracy": 0.9828102149069309,
+      "num_tokens": 17011175.0,
+      "step": 3200
+    },
+    {
+      "entropy": 0.0229463691044657,
+      "epoch": 0.3451926298157454,
+      "grad_norm": 3.9375,
+      "learning_rate": 1.5230682786379958e-05,
+      "loss": 0.0237918958067894,
+      "mean_token_accuracy": 0.9901813805103302,
+      "num_tokens": 17118025.0,
+      "step": 3220
+    },
+    {
+      "entropy": 0.02446629966070759,
+      "epoch": 0.3473366834170854,
+      "grad_norm": 1.3359375,
+      "learning_rate": 1.5171378245678398e-05,
+      "loss": 0.025769346952438356,
+      "mean_token_accuracy": 0.988752768561244,
+      "num_tokens": 17225787.0,
+      "step": 3240
+    },
+    {
+      "entropy": 0.023929561218301386,
+      "epoch": 0.34948073701842547,
+      "grad_norm": 2.890625,
+      "learning_rate": 1.5111824381531543e-05,
+      "loss": 0.02350829541683197,
+      "mean_token_accuracy": 0.9899691499769687,
+      "num_tokens": 17326334.0,
+      "step": 3260
+    },
+    {
+      "entropy": 0.023094384433488813,
+      "epoch": 0.3516247906197655,
+      "grad_norm": 1.3125,
+      "learning_rate": 1.5052024065161321e-05,
+      "loss": 0.025211021304130554,
+      "mean_token_accuracy": 0.989602592214942,
+      "num_tokens": 17427911.0,
+      "step": 3280
+    },
+    {
+      "entropy": 0.02580152012506005,
+      "epoch": 0.35376884422110555,
+      "grad_norm": 0.88671875,
+      "learning_rate": 1.4991980179671656e-05,
+      "loss": 0.028021180629730226,
+      "mean_token_accuracy": 0.9875154685229063,
+      "num_tokens": 17535492.0,
+      "step": 3300
+    },
+    {
+      "entropy": 0.025954745234002985,
+      "epoch": 0.35591289782244556,
+      "grad_norm": 1.46875,
+      "learning_rate": 1.4931695619909466e-05,
+      "loss": 0.023851674795150758,
+      "mean_token_accuracy": 0.9898436293005943,
+      "num_tokens": 17641670.0,
+      "step": 3320
+    },
+    {
+      "entropy": 0.024273743057074172,
+      "epoch": 0.3580569514237856,
+      "grad_norm": 1.7109375,
+      "learning_rate": 1.4871173292325106e-05,
+      "loss": 0.02511066496372223,
+      "mean_token_accuracy": 0.989355530589819,
+      "num_tokens": 17742555.0,
+      "step": 3340
+    },
+    {
+      "entropy": 0.02353285207400404,
+      "epoch": 0.36020100502512564,
+      "grad_norm": 1.4921875,
+      "learning_rate": 1.4810416114832236e-05,
+      "loss": 0.02515583038330078,
+      "mean_token_accuracy": 0.9906750213354826,
+      "num_tokens": 17848901.0,
+      "step": 3360
+    },
+    {
+      "entropy": 0.022243685176727014,
+      "epoch": 0.36234505862646565,
+      "grad_norm": 2.453125,
+      "learning_rate": 1.474942701666715e-05,
+      "loss": 0.02309694290161133,
+      "mean_token_accuracy": 0.9912830431014299,
+      "num_tokens": 17955457.0,
+      "step": 3380
+    },
+    {
+      "entropy": 0.022329559885474737,
+      "epoch": 0.3644891122278057,
+      "grad_norm": 1.6171875,
+      "learning_rate": 1.4688208938247535e-05,
+      "loss": 0.024044957756996155,
+      "mean_token_accuracy": 0.9913502413779497,
+      "num_tokens": 18054439.0,
+      "step": 3400
+    },
+    {
+      "entropy": 0.023860087489720173,
+      "epoch": 0.36663316582914574,
+      "grad_norm": 1.84375,
+      "learning_rate": 1.462676483103072e-05,
+      "loss": 0.023865839838981627,
+      "mean_token_accuracy": 0.9902952261269092,
+      "num_tokens": 18159108.0,
+      "step": 3420
+    },
+    {
+      "entropy": 0.021247414961362666,
+      "epoch": 0.36877721943048575,
+      "grad_norm": 1.8359375,
+      "learning_rate": 1.4565097657371378e-05,
+      "loss": 0.021402713656425477,
+      "mean_token_accuracy": 0.9921348176896572,
+      "num_tokens": 18268330.0,
+      "step": 3440
+    },
+    {
+      "entropy": 0.021685771951888456,
+      "epoch": 0.3709212730318258,
+      "grad_norm": 1.2109375,
+      "learning_rate": 1.4503210390378701e-05,
+      "loss": 0.026408180594444275,
+      "mean_token_accuracy": 0.9894901543855668,
+      "num_tokens": 18376382.0,
+      "step": 3460
+    },
+    {
+      "entropy": 0.025453939728959086,
+      "epoch": 0.37306532663316583,
+      "grad_norm": 2.09375,
+      "learning_rate": 1.444110601377306e-05,
+      "loss": 0.026739346981048583,
+      "mean_token_accuracy": 0.9884634926915169,
+      "num_tokens": 18489633.0,
+      "step": 3480
+    },
+    {
+      "entropy": 0.02501170370624095,
+      "epoch": 0.37520938023450584,
+      "grad_norm": 2.03125,
+      "learning_rate": 1.437878752174216e-05,
+      "loss": 0.02536824345588684,
+      "mean_token_accuracy": 0.9898253668099641,
+      "num_tokens": 18597861.0,
+      "step": 3500
+    },
+    {
+      "epoch": 0.37520938023450584,
+      "eval_entropy": 0.0244218321183871,
+      "eval_loss": 0.02227754145860672,
+      "eval_mean_token_accuracy": 0.9911262426376343,
+      "eval_num_tokens": 18597861.0,
+      "eval_runtime": 20.324,
+      "eval_samples_per_second": 36.902,
+      "eval_steps_per_second": 18.451,
+      "step": 3500
+    },
+    {
+      "entropy": 0.0246668197542931,
+      "epoch": 0.3773534338358459,
+      "grad_norm": 1.7421875,
+      "learning_rate": 1.4316257918796671e-05,
+      "loss": 0.025204175710678102,
+      "mean_token_accuracy": 0.99043660312891,
+      "num_tokens": 18697672.0,
+      "step": 3520
+    },
+    {
+      "entropy": 0.022950895108533587,
+      "epoch": 0.3794974874371859,
+      "grad_norm": 1.828125,
+      "learning_rate": 1.4253520219625387e-05,
+      "loss": 0.02428869903087616,
+      "mean_token_accuracy": 0.9907822337001562,
+      "num_tokens": 18811789.0,
+      "step": 3540
+    },
+    {
+      "entropy": 0.023854929314802577,
+      "epoch": 0.381641541038526,
+      "grad_norm": 1.9609375,
+      "learning_rate": 1.4190577448949876e-05,
+      "loss": 0.023869043588638304,
+      "mean_token_accuracy": 0.9913762111216784,
+      "num_tokens": 18911590.0,
+      "step": 3560
+    },
+    {
+      "entropy": 0.024182767125876126,
+      "epoch": 0.383785594639866,
+      "grad_norm": 2.546875,
+      "learning_rate": 1.4127432641378649e-05,
+      "loss": 0.023478889465332033,
+      "mean_token_accuracy": 0.9907276172190904,
+      "num_tokens": 19026734.0,
+      "step": 3580
+    },
+    {
+      "entropy": 0.020359245626377743,
+      "epoch": 0.385929648241206,
+      "grad_norm": 1.6640625,
+      "learning_rate": 1.4064088841260857e-05,
+      "loss": 0.023241102695465088,
+      "mean_token_accuracy": 0.9907599698752165,
+      "num_tokens": 19129417.0,
+      "step": 3600
+    },
+    {
+      "entropy": 0.021383443258810076,
+      "epoch": 0.3880737018425461,
+      "grad_norm": 1.3203125,
+      "learning_rate": 1.4000549102539516e-05,
+      "loss": 0.023419806361198427,
+      "mean_token_accuracy": 0.990417006239295,
+      "num_tokens": 19235205.0,
+      "step": 3620
+    },
+    {
+      "entropy": 0.020958786468827383,
+      "epoch": 0.3902177554438861,
+      "grad_norm": 1.046875,
+      "learning_rate": 1.3936816488604273e-05,
+      "loss": 0.02310475558042526,
+      "mean_token_accuracy": 0.991861392185092,
+      "num_tokens": 19333977.0,
+      "step": 3640
+    },
+    {
+      "entropy": 0.022316582887765436,
+      "epoch": 0.3923618090452261,
+      "grad_norm": 2.84375,
+      "learning_rate": 1.3872894072143708e-05,
+      "loss": 0.02316024452447891,
+      "mean_token_accuracy": 0.9916662760078907,
+      "num_tokens": 19433681.0,
+      "step": 3660
+    },
+    {
+      "entropy": 0.023413096193553428,
+      "epoch": 0.3945058626465662,
+      "grad_norm": 1.765625,
+      "learning_rate": 1.3808784934997199e-05,
+      "loss": 0.026311498880386353,
+      "mean_token_accuracy": 0.9898762315511703,
+      "num_tokens": 19543600.0,
+      "step": 3680
+    },
+    {
+      "entropy": 0.02430928884550667,
+      "epoch": 0.3966499162479062,
+      "grad_norm": 1.640625,
+      "learning_rate": 1.3744492168006332e-05,
+      "loss": 0.023604103922843934,
+      "mean_token_accuracy": 0.9905571412295103,
+      "num_tokens": 19656000.0,
+      "step": 3700
+    },
+    {
+      "entropy": 0.02260660900547009,
+      "epoch": 0.39879396984924625,
+      "grad_norm": 2.078125,
+      "learning_rate": 1.368001887086589e-05,
+      "loss": 0.024041768908500672,
+      "mean_token_accuracy": 0.9895115055143833,
+      "num_tokens": 19756250.0,
+      "step": 3720
+    },
+    {
+      "entropy": 0.02264130887056126,
+      "epoch": 0.40093802345058627,
+      "grad_norm": 1.625,
+      "learning_rate": 1.361536815197441e-05,
+      "loss": 0.025104725360870363,
+      "mean_token_accuracy": 0.9906443733721971,
+      "num_tokens": 19878359.0,
+      "step": 3740
+    },
+    {
+      "entropy": 0.023430266666673562,
+      "epoch": 0.4030820770519263,
+      "grad_norm": 1.65625,
+      "learning_rate": 1.3550543128284323e-05,
+      "loss": 0.024283410608768465,
+      "mean_token_accuracy": 0.9907973647117615,
+      "num_tokens": 19972202.0,
+      "step": 3760
+    },
+    {
+      "entropy": 0.02248989698346122,
+      "epoch": 0.40522613065326635,
+      "grad_norm": 1.4609375,
+      "learning_rate": 1.348554692515167e-05,
+      "loss": 0.02604263722896576,
+      "mean_token_accuracy": 0.9889465566724539,
+      "num_tokens": 20076482.0,
+      "step": 3780
+    },
+    {
+      "entropy": 0.02452128016102506,
+      "epoch": 0.40737018425460636,
+      "grad_norm": 1.703125,
+      "learning_rate": 1.3420382676185421e-05,
+      "loss": 0.025164645910263062,
+      "mean_token_accuracy": 0.9902698881924152,
+      "num_tokens": 20186450.0,
+      "step": 3800
+    },
+    {
+      "entropy": 0.024925458200004868,
+      "epoch": 0.4095142378559464,
+      "grad_norm": 1.921875,
+      "learning_rate": 1.335505352309642e-05,
+      "loss": 0.025616466999053955,
+      "mean_token_accuracy": 0.989701810106635,
+      "num_tokens": 20290995.0,
+      "step": 3820
+    },
+    {
+      "entropy": 0.022864083916329037,
+      "epoch": 0.41165829145728644,
+      "grad_norm": 1.8125,
+      "learning_rate": 1.3289562615545883e-05,
+      "loss": 0.02312360405921936,
+      "mean_token_accuracy": 0.9906341150403023,
+      "num_tokens": 20392839.0,
+      "step": 3840
+    },
+    {
+      "entropy": 0.022927879701637722,
+      "epoch": 0.41380234505862645,
+      "grad_norm": 1.2421875,
+      "learning_rate": 1.3223913110993576e-05,
+      "loss": 0.02285770922899246,
+      "mean_token_accuracy": 0.9910362754017115,
+      "num_tokens": 20499794.0,
+      "step": 3860
+    },
+    {
+      "entropy": 0.02117061285384807,
+      "epoch": 0.4159463986599665,
+      "grad_norm": 11.0,
+      "learning_rate": 1.3158108174545556e-05,
+      "loss": 0.021062937378883363,
+      "mean_token_accuracy": 0.9917857561260461,
+      "num_tokens": 20612104.0,
+      "step": 3880
+    },
+    {
+      "entropy": 0.02190584047739321,
+      "epoch": 0.41809045226130653,
+      "grad_norm": 2.625,
+      "learning_rate": 1.3092150978801605e-05,
+      "loss": 0.023377425968647003,
+      "mean_token_accuracy": 0.9904357828199863,
+      "num_tokens": 20723068.0,
+      "step": 3900
+    },
+    {
+      "entropy": 0.02153953154429473,
+      "epoch": 0.42023450586264655,
+      "grad_norm": 2.703125,
+      "learning_rate": 1.3026044703702249e-05,
+      "loss": 0.023467777669429778,
+      "mean_token_accuracy": 0.9905689381062984,
+      "num_tokens": 20826557.0,
+      "step": 3920
+    },
+    {
+      "entropy": 0.02152182651061594,
+      "epoch": 0.4223785594639866,
+      "grad_norm": 1.671875,
+      "learning_rate": 1.2959792536375471e-05,
+      "loss": 0.020291084051132204,
+      "mean_token_accuracy": 0.992015489563346,
+      "num_tokens": 20929814.0,
+      "step": 3940
+    },
+    {
+      "entropy": 0.02025915689789599,
+      "epoch": 0.4245226130653266,
+      "grad_norm": 1.7421875,
+      "learning_rate": 1.2893397670983022e-05,
+      "loss": 0.020092053711414336,
+      "mean_token_accuracy": 0.9926223222166299,
+      "num_tokens": 21032536.0,
+      "step": 3960
+    },
+    {
+      "entropy": 0.019897875330707394,
+      "epoch": 0.4266666666666667,
+      "grad_norm": 2.46875,
+      "learning_rate": 1.2826863308566437e-05,
+      "loss": 0.022923128306865694,
+      "mean_token_accuracy": 0.9915768329054118,
+      "num_tokens": 21125010.0,
+      "step": 3980
+    },
+    {
+      "entropy": 0.020706827863295984,
+      "epoch": 0.4288107202680067,
+      "grad_norm": 1.9140625,
+      "learning_rate": 1.2760192656892723e-05,
+      "loss": 0.022419747710227967,
+      "mean_token_accuracy": 0.9913075774908066,
+      "num_tokens": 21242867.0,
+      "step": 4000
+    },
+    {
+      "epoch": 0.4288107202680067,
+      "eval_entropy": 0.01964530904859809,
+      "eval_loss": 0.02178012579679489,
+      "eval_mean_token_accuracy": 0.9911314179102579,
+      "eval_num_tokens": 21242867.0,
+      "eval_runtime": 20.3146,
+      "eval_samples_per_second": 36.919,
+      "eval_steps_per_second": 18.46,
+      "step": 4000
+    },
+    {
+      "entropy": 0.02013387772437909,
+      "epoch": 0.4309547738693467,
+      "grad_norm": 1.5078125,
+      "learning_rate": 1.2693388930299672e-05,
+      "loss": 0.021522065997123717,
+      "mean_token_accuracy": 0.9929842289537192,
+      "num_tokens": 21346530.0,
+      "step": 4020
+    },
+    {
+      "entropy": 0.0202242263831522,
+      "epoch": 0.4330988274706868,
+      "grad_norm": 2.828125,
+      "learning_rate": 1.2626455349540918e-05,
+      "loss": 0.022661156952381134,
+      "mean_token_accuracy": 0.9919143192470073,
+      "num_tokens": 21458250.0,
+      "step": 4040
+    },
+    {
+      "entropy": 0.022103192698341444,
+      "epoch": 0.4352428810720268,
+      "grad_norm": 1.3671875,
+      "learning_rate": 1.2559395141630656e-05,
+      "loss": 0.02430972158908844,
+      "mean_token_accuracy": 0.9904460556805134,
+      "num_tokens": 21563008.0,
+      "step": 4060
+    },
+    {
+      "entropy": 0.02260455262266987,
+      "epoch": 0.4373869346733668,
+      "grad_norm": 1.4609375,
+      "learning_rate": 1.2492211539688036e-05,
+      "loss": 0.025785067677497865,
+      "mean_token_accuracy": 0.9895172834396362,
+      "num_tokens": 21668155.0,
+      "step": 4080
+    },
+    {
+      "entropy": 0.02297381363787281,
+      "epoch": 0.4395309882747069,
+      "grad_norm": 2.203125,
+      "learning_rate": 1.2424907782781313e-05,
+      "loss": 0.024609476327896118,
+      "mean_token_accuracy": 0.990400804951787,
+      "num_tokens": 21784031.0,
+      "step": 4100
+    },
+    {
+      "entropy": 0.022381592743295188,
+      "epoch": 0.4416750418760469,
+      "grad_norm": 1.609375,
+      "learning_rate": 1.2357487115771673e-05,
+      "loss": 0.023763097822666168,
+      "mean_token_accuracy": 0.99109130166471,
+      "num_tokens": 21893847.0,
+      "step": 4120
+    },
+    {
+      "entropy": 0.022086531718105106,
+      "epoch": 0.44381909547738696,
+      "grad_norm": 2.0,
+      "learning_rate": 1.2289952789156798e-05,
+      "loss": 0.020330893993377685,
+      "mean_token_accuracy": 0.9921020328998565,
+      "num_tokens": 22005454.0,
+      "step": 4140
+    },
+    {
+      "entropy": 0.02248057241056358,
+      "epoch": 0.445963149078727,
+      "grad_norm": 1.609375,
+      "learning_rate": 1.222230805891414e-05,
+      "loss": 0.02294687330722809,
+      "mean_token_accuracy": 0.9910975113511086,
+      "num_tokens": 22107232.0,
+      "step": 4160
+    },
+    {
+      "entropy": 0.02265706049165601,
+      "epoch": 0.448107202680067,
+      "grad_norm": 3.90625,
+      "learning_rate": 1.215455618634396e-05,
+      "loss": 0.025840884447097777,
+      "mean_token_accuracy": 0.9897211868315935,
+      "num_tokens": 22217548.0,
+      "step": 4180
+    },
+    {
+      "entropy": 0.02281195720179312,
+      "epoch": 0.45025125628140705,
+      "grad_norm": 3.1875,
+      "learning_rate": 1.2086700437912077e-05,
+      "loss": 0.024470710754394533,
+      "mean_token_accuracy": 0.9903297979384661,
+      "num_tokens": 22328036.0,
+      "step": 4200
+    },
+    {
+      "entropy": 0.02147264582108619,
+      "epoch": 0.45239530988274707,
+      "grad_norm": 1.6328125,
+      "learning_rate": 1.2018744085092399e-05,
+      "loss": 0.021890904009342193,
+      "mean_token_accuracy": 0.9921745974570513,
+      "num_tokens": 22439360.0,
+      "step": 4220
+    },
+    {
+      "entropy": 0.021090176403095028,
+      "epoch": 0.4545393634840871,
+      "grad_norm": 1.6171875,
+      "learning_rate": 1.1950690404209186e-05,
+      "loss": 0.022462236881256103,
+      "mean_token_accuracy": 0.992034525051713,
+      "num_tokens": 22543357.0,
+      "step": 4240
+    },
+    {
+      "entropy": 0.022680416493108167,
+      "epoch": 0.45668341708542715,
+      "grad_norm": 1.328125,
+      "learning_rate": 1.1882542676279098e-05,
+      "loss": 0.02402680218219757,
+      "mean_token_accuracy": 0.9908802334219218,
+      "num_tokens": 22650751.0,
+      "step": 4260
+    },
+    {
+      "entropy": 0.0232969627111288,
+      "epoch": 0.45882747068676716,
+      "grad_norm": 2.96875,
+      "learning_rate": 1.1814304186853009e-05,
+      "loss": 0.022290503978729247,
+      "mean_token_accuracy": 0.991101274266839,
+      "num_tokens": 22753548.0,
+      "step": 4280
+    },
+    {
+      "entropy": 0.021912004817795604,
+      "epoch": 0.4609715242881072,
+      "grad_norm": 3.359375,
+      "learning_rate": 1.1745978225857609e-05,
+      "loss": 0.022448597848415373,
+      "mean_token_accuracy": 0.9917845461517573,
+      "num_tokens": 22861310.0,
+      "step": 4300
+    },
+    {
+      "entropy": 0.021719314875508645,
+      "epoch": 0.46311557788944724,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.1677568087436767e-05,
+      "loss": 0.023140157759189605,
+      "mean_token_accuracy": 0.9908079355955124,
+      "num_tokens": 22978927.0,
+      "step": 4320
+    },
+    {
+      "entropy": 0.021950446777191247,
+      "epoch": 0.46525963149078725,
+      "grad_norm": 2.203125,
+      "learning_rate": 1.1609077069792754e-05,
+      "loss": 0.0213756799697876,
+      "mean_token_accuracy": 0.9913494195789099,
+      "num_tokens": 23082226.0,
+      "step": 4340
+    },
+    {
+      "entropy": 0.02140526709590631,
+      "epoch": 0.4674036850921273,
+      "grad_norm": 2.15625,
+      "learning_rate": 1.1540508475027179e-05,
+      "loss": 0.020155489444732666,
+      "mean_token_accuracy": 0.9919397257268429,
+      "num_tokens": 23180655.0,
+      "step": 4360
+    },
+    {
+      "entropy": 0.017593615273926843,
+      "epoch": 0.46954773869346733,
+      "grad_norm": 1.75,
+      "learning_rate": 1.147186560898183e-05,
+      "loss": 0.02236589938402176,
+      "mean_token_accuracy": 0.9917930167168378,
+      "num_tokens": 23283721.0,
+      "step": 4380
+    },
+    {
+      "entropy": 0.020877945362190075,
+      "epoch": 0.47169179229480734,
+      "grad_norm": 1.625,
+      "learning_rate": 1.1403151781079278e-05,
+      "loss": 0.022862054407596588,
+      "mean_token_accuracy": 0.9906388968229294,
+      "num_tokens": 23390288.0,
+      "step": 4400
+    },
+    {
+      "entropy": 0.01995794570948419,
+      "epoch": 0.4738358458961474,
+      "grad_norm": 2.109375,
+      "learning_rate": 1.1334370304163306e-05,
+      "loss": 0.02168610394001007,
+      "mean_token_accuracy": 0.9925977524369956,
+      "num_tokens": 23485464.0,
+      "step": 4420
+    },
+    {
+      "entropy": 0.01932002087028195,
+      "epoch": 0.4759798994974874,
+      "grad_norm": 1.6796875,
+      "learning_rate": 1.1265524494339219e-05,
+      "loss": 0.02263712286949158,
+      "mean_token_accuracy": 0.9910555865615607,
+      "num_tokens": 23585708.0,
+      "step": 4440
+    },
+    {
+      "entropy": 0.021627131735112926,
+      "epoch": 0.4781239530988275,
+      "grad_norm": 2.140625,
+      "learning_rate": 1.1196617670813936e-05,
+      "loss": 0.022156818211078642,
+      "mean_token_accuracy": 0.9912719707936049,
+      "num_tokens": 23693455.0,
+      "step": 4460
+    },
+    {
+      "entropy": 0.02139281702184235,
+      "epoch": 0.4802680067001675,
+      "grad_norm": 2.1875,
+      "learning_rate": 1.1127653155735993e-05,
+      "loss": 0.022481553256511688,
+      "mean_token_accuracy": 0.9902749702334404,
+      "num_tokens": 23800058.0,
+      "step": 4480
+    },
+    {
+      "entropy": 0.02091277845497643,
+      "epoch": 0.4824120603015075,
+      "grad_norm": 2.328125,
+      "learning_rate": 1.1058634274035361e-05,
+      "loss": 0.02205648124217987,
+      "mean_token_accuracy": 0.9916612818837166,
+      "num_tokens": 23910439.0,
+      "step": 4500
+    },
+    {
+      "epoch": 0.4824120603015075,
+      "eval_entropy": 0.021846828331123107,
+      "eval_loss": 0.02092284895479679,
+      "eval_mean_token_accuracy": 0.9913039420445761,
+      "eval_num_tokens": 23910439.0,
+      "eval_runtime": 20.2669,
+      "eval_samples_per_second": 37.006,
+      "eval_steps_per_second": 18.503,
+      "step": 4500
+    },
+    {
+      "entropy": 0.021989129094345118,
+      "epoch": 0.4845561139028476,
+      "grad_norm": 1.765625,
+      "learning_rate": 1.0989564353263133e-05,
+      "loss": 0.02231614738702774,
+      "mean_token_accuracy": 0.9905249148607254,
+      "num_tokens": 24019503.0,
+      "step": 4520
+    },
+    {
+      "entropy": 0.02183805209115235,
+      "epoch": 0.4867001675041876,
+      "grad_norm": 2.3125,
+      "learning_rate": 1.0920446723431127e-05,
+      "loss": 0.02185622453689575,
+      "mean_token_accuracy": 0.9918582946062088,
+      "num_tokens": 24124480.0,
+      "step": 4540
+    },
+    {
+      "entropy": 0.02131056402449758,
+      "epoch": 0.48884422110552767,
+      "grad_norm": 1.3203125,
+      "learning_rate": 1.0851284716851295e-05,
+      "loss": 0.022905312478542328,
+      "mean_token_accuracy": 0.9910277899354696,
+      "num_tokens": 24230705.0,
+      "step": 4560
+    },
+    {
+      "entropy": 0.023312085462430333,
+      "epoch": 0.4909882747068677,
+      "grad_norm": 1.8828125,
+      "learning_rate": 1.0782081667975106e-05,
+      "loss": 0.022799497842788695,
+      "mean_token_accuracy": 0.9904710713773965,
+      "num_tokens": 24335812.0,
+      "step": 4580
+    },
+    {
+      "entropy": 0.02137677239184086,
+      "epoch": 0.4931323283082077,
+      "grad_norm": 1.34375,
+      "learning_rate": 1.0712840913232768e-05,
+      "loss": 0.02341741919517517,
+      "mean_token_accuracy": 0.9909932963550091,
+      "num_tokens": 24444839.0,
+      "step": 4600
+    },
+    {
+      "entropy": 0.022716682115378715,
+      "epoch": 0.49527638190954776,
+      "grad_norm": 1.484375,
+      "learning_rate": 1.0643565790872359e-05,
+      "loss": 0.02410309910774231,
+      "mean_token_accuracy": 0.9909039065241814,
+      "num_tokens": 24545937.0,
+      "step": 4620
+    },
+    {
+      "entropy": 0.022203819393803316,
+      "epoch": 0.49742043551088777,
+      "grad_norm": 1.8515625,
+      "learning_rate": 1.0574259640798908e-05,
+      "loss": 0.022010770440101624,
+      "mean_token_accuracy": 0.9920209400355816,
+      "num_tokens": 24647974.0,
+      "step": 4640
+    },
+    {
+      "entropy": 0.02109122319297967,
+      "epoch": 0.4995644891122278,
+      "grad_norm": 1.3359375,
+      "learning_rate": 1.0504925804413348e-05,
+      "loss": 0.02280532568693161,
+      "mean_token_accuracy": 0.9912972040474415,
+      "num_tokens": 24752530.0,
+      "step": 4660
+    },
+    {
+      "entropy": 0.022384429502744753,
+      "epoch": 0.5017085427135678,
+      "grad_norm": 2.390625,
+      "learning_rate": 1.0435567624451435e-05,
+      "loss": 0.027354520559310914,
+      "mean_token_accuracy": 0.98953907340765,
+      "num_tokens": 24854708.0,
+      "step": 4680
+    },
+    {
+      "entropy": 0.02178340100826972,
+      "epoch": 0.5038525963149079,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.0366188444822577e-05,
+      "loss": 0.02108224630355835,
+      "mean_token_accuracy": 0.991971093788743,
+      "num_tokens": 24951913.0,
+      "step": 4700
+    },
+    {
+      "entropy": 0.0228582585350523,
+      "epoch": 0.5059966499162479,
+      "grad_norm": 1.578125,
+      "learning_rate": 1.0296791610448634e-05,
+      "loss": 0.02327609956264496,
+      "mean_token_accuracy": 0.9911777719855308,
+      "num_tokens": 25057377.0,
+      "step": 4720
+    },
+    {
+      "entropy": 0.022045019802681053,
+      "epoch": 0.5081407035175879,
+      "grad_norm": 1.1953125,
+      "learning_rate": 1.0227380467102627e-05,
+      "loss": 0.023243658244609833,
+      "mean_token_accuracy": 0.9909518156200647,
+      "num_tokens": 25170778.0,
+      "step": 4740
+    },
+    {
+      "entropy": 0.01989219032784604,
+      "epoch": 0.510284757118928,
+      "grad_norm": 5.03125,
+      "learning_rate": 1.0157958361247448e-05,
+      "loss": 0.020330929756164552,
+      "mean_token_accuracy": 0.9921452704817056,
+      "num_tokens": 25284310.0,
+      "step": 4760
+    },
+    {
+      "entropy": 0.021722077413278383,
+      "epoch": 0.512428810720268,
+      "grad_norm": 1.6015625,
+      "learning_rate": 1.0088528639874516e-05,
+      "loss": 0.02167322486639023,
+      "mean_token_accuracy": 0.9913335714489222,
+      "num_tokens": 25392904.0,
+      "step": 4780
+    },
+    {
+      "entropy": 0.022090970205681515,
+      "epoch": 0.5145728643216081,
+      "grad_norm": 1.390625,
+      "learning_rate": 1.0019094650342409e-05,
+      "loss": 0.023602089285850524,
+      "mean_token_accuracy": 0.9912930846214294,
+      "num_tokens": 25505548.0,
+      "step": 4800
+    },
+    {
+      "entropy": 0.021318517662666635,
+      "epoch": 0.516716917922948,
+      "grad_norm": 1.8046875,
+      "learning_rate": 9.94965974021548e-06,
+      "loss": 0.021860493719577788,
+      "mean_token_accuracy": 0.9906282421201468,
+      "num_tokens": 25616943.0,
+      "step": 4820
+    },
+    {
+      "entropy": 0.020716092844122614,
+      "epoch": 0.5188609715242881,
+      "grad_norm": 1.7109375,
+      "learning_rate": 9.880227257102483e-06,
+      "loss": 0.021078214049339294,
+      "mean_token_accuracy": 0.9921130206435919,
+      "num_tokens": 25727510.0,
+      "step": 4840
+    },
+    {
+      "entropy": 0.02081386552129061,
+      "epoch": 0.5210050251256282,
+      "grad_norm": 2.203125,
+      "learning_rate": 9.810800548495132e-06,
+      "loss": 0.019788502156734465,
+      "mean_token_accuracy": 0.9920309089124203,
+      "num_tokens": 25831138.0,
+      "step": 4860
+    },
+    {
+      "entropy": 0.020538337337939083,
+      "epoch": 0.5231490787269681,
+      "grad_norm": 3.671875,
+      "learning_rate": 9.741382961606768e-06,
+      "loss": 0.021504290401935577,
+      "mean_token_accuracy": 0.9916285052895546,
+      "num_tokens": 25938256.0,
+      "step": 4880
+    },
+    {
+      "entropy": 0.02026532054192103,
+      "epoch": 0.5252931323283082,
+      "grad_norm": 2.015625,
+      "learning_rate": 9.671977843210928e-06,
+      "loss": 0.022037246823310853,
+      "mean_token_accuracy": 0.9909355752170086,
+      "num_tokens": 26045653.0,
+      "step": 4900
+    },
+    {
+      "entropy": 0.021552794340914262,
+      "epoch": 0.5274371859296483,
+      "grad_norm": 1.9453125,
+      "learning_rate": 9.602588539480037e-06,
+      "loss": 0.024588188529014586,
+      "mean_token_accuracy": 0.9902191430330276,
+      "num_tokens": 26147507.0,
+      "step": 4920
+    },
+    {
+      "entropy": 0.02167862878472988,
+      "epoch": 0.5295812395309882,
+      "grad_norm": 1.9921875,
+      "learning_rate": 9.533218395824046e-06,
+      "loss": 0.023571619391441347,
+      "mean_token_accuracy": 0.9900527201592922,
+      "num_tokens": 26257125.0,
+      "step": 4940
+    },
+    {
+      "entropy": 0.02021576279623787,
+      "epoch": 0.5317252931323283,
+      "grad_norm": 2.1875,
+      "learning_rate": 9.463870756729163e-06,
+      "loss": 0.019122126698493957,
+      "mean_token_accuracy": 0.992604573443532,
+      "num_tokens": 26364694.0,
+      "step": 4960
+    },
+    {
+      "entropy": 0.019900748710369953,
+      "epoch": 0.5338693467336684,
+      "grad_norm": 2.0625,
+      "learning_rate": 9.394548965596609e-06,
+      "loss": 0.01807314157485962,
+      "mean_token_accuracy": 0.9929017867892981,
+      "num_tokens": 26473983.0,
+      "step": 4980
+    },
+    {
+      "entropy": 0.0207442283473938,
+      "epoch": 0.5360134003350083,
+      "grad_norm": 1.4375,
+      "learning_rate": 9.325256364581403e-06,
+      "loss": 0.022004596889019012,
+      "mean_token_accuracy": 0.9910150520503521,
+      "num_tokens": 26586291.0,
+      "step": 5000
+    },
+    {
+      "epoch": 0.5360134003350083,
+      "eval_entropy": 0.020443486923235468,
+      "eval_loss": 0.01984340138733387,
+      "eval_mean_token_accuracy": 0.9918238368034363,
+      "eval_num_tokens": 26586291.0,
+      "eval_runtime": 20.253,
+      "eval_samples_per_second": 37.032,
+      "eval_steps_per_second": 18.516,
+      "step": 5000
+    },
+    {
+      "entropy": 0.02020404681409218,
+      "epoch": 0.5381574539363484,
+      "grad_norm": 2.546875,
+      "learning_rate": 9.255996294431251e-06,
+      "loss": 0.018390218913555145,
+      "mean_token_accuracy": 0.992999167367816,
+      "num_tokens": 26704783.0,
+      "step": 5020
+    },
+    {
+      "entropy": 0.01915527433520765,
+      "epoch": 0.5403015075376885,
+      "grad_norm": 2.078125,
+      "learning_rate": 9.186772094325482e-06,
+      "loss": 0.02050086110830307,
+      "mean_token_accuracy": 0.991879815980792,
+      "num_tokens": 26801816.0,
+      "step": 5040
+    },
+    {
+      "entropy": 0.01799158312455802,
+      "epoch": 0.5424455611390284,
+      "grad_norm": 1.828125,
+      "learning_rate": 9.117587101714046e-06,
+      "loss": 0.02155505567789078,
+      "mean_token_accuracy": 0.9908429838716983,
+      "num_tokens": 26908853.0,
+      "step": 5060
+    },
+    {
+      "entropy": 0.020375365777476874,
+      "epoch": 0.5445896147403685,
+      "grad_norm": 1.75,
+      "learning_rate": 9.048444652156607e-06,
+      "loss": 0.022749082744121553,
+      "mean_token_accuracy": 0.991591277718544,
+      "num_tokens": 27027467.0,
+      "step": 5080
+    },
+    {
+      "entropy": 0.01897250141009863,
+      "epoch": 0.5467336683417086,
+      "grad_norm": 1.9375,
+      "learning_rate": 8.979348079161753e-06,
+      "loss": 0.021764059364795686,
+      "mean_token_accuracy": 0.9915874369442463,
+      "num_tokens": 27140253.0,
+      "step": 5100
+    },
+    {
+      "entropy": 0.020416706607784364,
+      "epoch": 0.5488777219430486,
+      "grad_norm": 1.515625,
+      "learning_rate": 8.910300714026242e-06,
+      "loss": 0.019444821774959563,
+      "mean_token_accuracy": 0.9921912178397179,
+      "num_tokens": 27250990.0,
+      "step": 5120
+    },
+    {
+      "entropy": 0.019052070355746763,
+      "epoch": 0.5510217755443886,
+      "grad_norm": 0.79296875,
+      "learning_rate": 8.841305885674431e-06,
+      "loss": 0.0189539834856987,
+      "mean_token_accuracy": 0.994080551713705,
+      "num_tokens": 27355412.0,
+      "step": 5140
+    },
+    {
+      "entropy": 0.019072407830481097,
+      "epoch": 0.5531658291457286,
+      "grad_norm": 2.125,
+      "learning_rate": 8.772366920497754e-06,
+      "loss": 0.022655683755874633,
+      "mean_token_accuracy": 0.9920244321227074,
+      "num_tokens": 27451798.0,
+      "step": 5160
+    },
+    {
+      "entropy": 0.0204707554947845,
+      "epoch": 0.5553098827470687,
+      "grad_norm": 1.765625,
+      "learning_rate": 8.70348714219436e-06,
+      "loss": 0.02314187288284302,
+      "mean_token_accuracy": 0.9906669486314058,
+      "num_tokens": 27566715.0,
+      "step": 5180
+    },
+    {
+      "entropy": 0.01811999880219446,
+      "epoch": 0.5574539363484087,
+      "grad_norm": 1.375,
+      "learning_rate": 8.634669871608879e-06,
+      "loss": 0.021425962448120117,
+      "mean_token_accuracy": 0.9925322774797678,
+      "num_tokens": 27659771.0,
+      "step": 5200
+    },
+    {
+      "entropy": 0.02080239012641414,
+      "epoch": 0.5595979899497487,
+      "grad_norm": 1.921875,
+      "learning_rate": 8.565918426572287e-06,
+      "loss": 0.021234531700611115,
+      "mean_token_accuracy": 0.9914034195244312,
+      "num_tokens": 27774261.0,
+      "step": 5220
+    },
+    {
+      "entropy": 0.018487965166877986,
+      "epoch": 0.5617420435510888,
+      "grad_norm": 1.7421875,
+      "learning_rate": 8.497236121741988e-06,
+      "loss": 0.020122981071472167,
+      "mean_token_accuracy": 0.9918253514915705,
+      "num_tokens": 27873405.0,
+      "step": 5240
+    },
+    {
+      "entropy": 0.020079654125311208,
+      "epoch": 0.5638860971524288,
+      "grad_norm": 1.875,
+      "learning_rate": 8.42862626844198e-06,
+      "loss": 0.0222967728972435,
+      "mean_token_accuracy": 0.9918321210891008,
+      "num_tokens": 27980906.0,
+      "step": 5260
+    },
+    {
+      "entropy": 0.02020533189497655,
+      "epoch": 0.5660301507537688,
+      "grad_norm": 2.046875,
+      "learning_rate": 8.3600921745032e-06,
+      "loss": 0.02204359769821167,
+      "mean_token_accuracy": 0.9903446055948735,
+      "num_tokens": 28085101.0,
+      "step": 5280
+    },
+    {
+      "entropy": 0.019518600214541947,
+      "epoch": 0.5681742043551089,
+      "grad_norm": 1.140625,
+      "learning_rate": 8.291637144104084e-06,
+      "loss": 0.02309599071741104,
+      "mean_token_accuracy": 0.9912882562726736,
+      "num_tokens": 28193879.0,
+      "step": 5300
+    },
+    {
+      "entropy": 0.02015163512619438,
+      "epoch": 0.5703182579564489,
+      "grad_norm": 1.828125,
+      "learning_rate": 8.223264477611218e-06,
+      "loss": 0.020999400317668913,
+      "mean_token_accuracy": 0.9915294233709574,
+      "num_tokens": 28296254.0,
+      "step": 5320
+    },
+    {
+      "entropy": 0.02264108727913481,
+      "epoch": 0.5724623115577889,
+      "grad_norm": 1.609375,
+      "learning_rate": 8.154977471420266e-06,
+      "loss": 0.02454686313867569,
+      "mean_token_accuracy": 0.9913088951259852,
+      "num_tokens": 28409744.0,
+      "step": 5340
+    },
+    {
+      "entropy": 0.02133122851785174,
+      "epoch": 0.574606365159129,
+      "grad_norm": 1.078125,
+      "learning_rate": 8.086779417797005e-06,
+      "loss": 0.018969587981700897,
+      "mean_token_accuracy": 0.9918996881693601,
+      "num_tokens": 28523944.0,
+      "step": 5360
+    },
+    {
+      "entropy": 0.020475312656594725,
+      "epoch": 0.5767504187604691,
+      "grad_norm": 2.578125,
+      "learning_rate": 8.018673604718621e-06,
+      "loss": 0.020480218529701232,
+      "mean_token_accuracy": 0.9909398626536131,
+      "num_tokens": 28620491.0,
+      "step": 5380
+    },
+    {
+      "entropy": 0.02034863719095483,
+      "epoch": 0.578894472361809,
+      "grad_norm": 2.3125,
+      "learning_rate": 7.950663315715189e-06,
+      "loss": 0.020605732500553132,
+      "mean_token_accuracy": 0.991118798777461,
+      "num_tokens": 28724436.0,
+      "step": 5400
+    },
+    {
+      "entropy": 0.019775752684540748,
+      "epoch": 0.5810385259631491,
+      "grad_norm": 3.78125,
+      "learning_rate": 7.882751829711343e-06,
+      "loss": 0.017187023162841798,
+      "mean_token_accuracy": 0.9925138726830482,
+      "num_tokens": 28834817.0,
+      "step": 5420
+    },
+    {
+      "entropy": 0.020291093751802693,
+      "epoch": 0.5831825795644892,
+      "grad_norm": 2.59375,
+      "learning_rate": 7.814942420868225e-06,
+      "loss": 0.02155611515045166,
+      "mean_token_accuracy": 0.9925528045743703,
+      "num_tokens": 28949852.0,
+      "step": 5440
+    },
+    {
+      "entropy": 0.020088849540070442,
+      "epoch": 0.5853266331658291,
+      "grad_norm": 1.4765625,
+      "learning_rate": 7.747238358425603e-06,
+      "loss": 0.021509405970573426,
+      "mean_token_accuracy": 0.9916152823716402,
+      "num_tokens": 29053126.0,
+      "step": 5460
+    },
+    {
+      "entropy": 0.02078960820804241,
+      "epoch": 0.5874706867671692,
+      "grad_norm": 1.6953125,
+      "learning_rate": 7.679642906544268e-06,
+      "loss": 0.022637182474136354,
+      "mean_token_accuracy": 0.9910664647817612,
+      "num_tokens": 29173776.0,
+      "step": 5480
+    },
+    {
+      "entropy": 0.019441102636301368,
+      "epoch": 0.5896147403685092,
+      "grad_norm": 3.515625,
+      "learning_rate": 7.6121593241486644e-06,
+      "loss": 0.021338345110416414,
+      "mean_token_accuracy": 0.9917159400880337,
+      "num_tokens": 29278255.0,
+      "step": 5500
+    },
+    {
+      "epoch": 0.5896147403685092,
+      "eval_entropy": 0.019579512438397312,
+      "eval_loss": 0.019862476736307144,
+      "eval_mean_token_accuracy": 0.9916693598429362,
+      "eval_num_tokens": 29278255.0,
+      "eval_runtime": 20.2793,
+      "eval_samples_per_second": 36.984,
+      "eval_steps_per_second": 18.492,
+      "step": 5500
+    },
+    {
+      "entropy": 0.01876566273849676,
+      "epoch": 0.5917587939698492,
+      "grad_norm": 1.53125,
+      "learning_rate": 7.5447908647697506e-06,
+      "loss": 0.02028908282518387,
+      "mean_token_accuracy": 0.9926750350743532,
+      "num_tokens": 29377393.0,
+      "step": 5520
+    },
+    {
+      "entropy": 0.019120201245004863,
+      "epoch": 0.5939028475711893,
+      "grad_norm": 2.09375,
+      "learning_rate": 7.47754077638817e-06,
+      "loss": 0.02107948064804077,
+      "mean_token_accuracy": 0.9924666147679091,
+      "num_tokens": 29481829.0,
+      "step": 5540
+    },
+    {
+      "entropy": 0.019154090119650392,
+      "epoch": 0.5960469011725293,
+      "grad_norm": 1.65625,
+      "learning_rate": 7.410412301277628e-06,
+      "loss": 0.021442399919033052,
+      "mean_token_accuracy": 0.9915456175804138,
+      "num_tokens": 29577548.0,
+      "step": 5560
+    },
+    {
+      "entropy": 0.020240560936554176,
+      "epoch": 0.5981909547738693,
+      "grad_norm": 1.71875,
+      "learning_rate": 7.343408675848593e-06,
+      "loss": 0.022295641899108886,
+      "mean_token_accuracy": 0.9908768083900213,
+      "num_tokens": 29693831.0,
+      "step": 5580
+    },
+    {
+      "entropy": 0.018894769826965786,
+      "epoch": 0.6003350083752094,
+      "grad_norm": 2.203125,
+      "learning_rate": 7.27653313049226e-06,
+      "loss": 0.017608430981636048,
+      "mean_token_accuracy": 0.9934392455965281,
+      "num_tokens": 29807626.0,
+      "step": 5600
+    },
+    {
+      "entropy": 0.018964907280769695,
+      "epoch": 0.6024790619765494,
+      "grad_norm": 1.7890625,
+      "learning_rate": 7.209788889424797e-06,
+      "loss": 0.021428127586841584,
+      "mean_token_accuracy": 0.992251905053854,
+      "num_tokens": 29918699.0,
+      "step": 5620
+    },
+    {
+      "entropy": 0.019173358734906287,
+      "epoch": 0.6046231155778894,
+      "grad_norm": 2.03125,
+      "learning_rate": 7.143179170531917e-06,
+      "loss": 0.02228628993034363,
+      "mean_token_accuracy": 0.9917910866439342,
+      "num_tokens": 30012605.0,
+      "step": 5640
+    },
+    {
+      "entropy": 0.020879295714121327,
+      "epoch": 0.6067671691792295,
+      "grad_norm": 1.71875,
+      "learning_rate": 7.076707185213716e-06,
+      "loss": 0.023567219078540803,
+      "mean_token_accuracy": 0.9915827866643667,
+      "num_tokens": 30121162.0,
+      "step": 5660
+    },
+    {
+      "entropy": 0.021870423488599043,
+      "epoch": 0.6089112227805695,
+      "grad_norm": 2.09375,
+      "learning_rate": 7.010376138229848e-06,
+      "loss": 0.023693473637104036,
+      "mean_token_accuracy": 0.9895996507257223,
+      "num_tokens": 30224756.0,
+      "step": 5680
+    },
+    {
+      "entropy": 0.02007417911304401,
+      "epoch": 0.6110552763819096,
+      "grad_norm": 1.6484375,
+      "learning_rate": 6.944189227545035e-06,
+      "loss": 0.018479639291763307,
+      "mean_token_accuracy": 0.992578137665987,
+      "num_tokens": 30328568.0,
+      "step": 5700
+    },
+    {
+      "entropy": 0.02106366403361335,
+      "epoch": 0.6131993299832496,
+      "grad_norm": 3.3125,
+      "learning_rate": 6.878149644174868e-06,
+      "loss": 0.021388374269008636,
+      "mean_token_accuracy": 0.9912810981273651,
+      "num_tokens": 30433800.0,
+      "step": 5720
+    },
+    {
+      "entropy": 0.019824264124895307,
+      "epoch": 0.6153433835845896,
+      "grad_norm": 3.828125,
+      "learning_rate": 6.812260572031966e-06,
+      "loss": 0.019826412200927734,
+      "mean_token_accuracy": 0.9915285713970661,
+      "num_tokens": 30533876.0,
+      "step": 5740
+    },
+    {
+      "entropy": 0.019220966260763816,
+      "epoch": 0.6174874371859297,
+      "grad_norm": 2.3125,
+      "learning_rate": 6.74652518777247e-06,
+      "loss": 0.023332512378692626,
+      "mean_token_accuracy": 0.9904117193073034,
+      "num_tokens": 30640243.0,
+      "step": 5760
+    },
+    {
+      "entropy": 0.019488869791348406,
+      "epoch": 0.6196314907872696,
+      "grad_norm": 1.7734375,
+      "learning_rate": 6.680946660642893e-06,
+      "loss": 0.023433607816696168,
+      "mean_token_accuracy": 0.991485184058547,
+      "num_tokens": 30740058.0,
+      "step": 5780
+    },
+    {
+      "entropy": 0.02283499206282613,
+      "epoch": 0.6217755443886097,
+      "grad_norm": 1.9296875,
+      "learning_rate": 6.615528152327328e-06,
+      "loss": 0.024059185385704042,
+      "mean_token_accuracy": 0.9913460418581963,
+      "num_tokens": 30852207.0,
+      "step": 5800
+    },
+    {
+      "entropy": 0.020180349141332952,
+      "epoch": 0.6239195979899498,
+      "grad_norm": 3.109375,
+      "learning_rate": 6.550272816795006e-06,
+      "loss": 0.01820368319749832,
+      "mean_token_accuracy": 0.9933834221214056,
+      "num_tokens": 30953138.0,
+      "step": 5820
+    },
+    {
+      "entropy": 0.019724707261821094,
+      "epoch": 0.6260636515912897,
+      "grad_norm": 1.65625,
+      "learning_rate": 6.485183800148248e-06,
+      "loss": 0.020257289707660674,
+      "mean_token_accuracy": 0.9930203214287758,
+      "num_tokens": 31064272.0,
+      "step": 5840
+    },
+    {
+      "entropy": 0.01873356030444029,
+      "epoch": 0.6282077051926298,
+      "grad_norm": 3.484375,
+      "learning_rate": 6.420264240470771e-06,
+      "loss": 0.02272033840417862,
+      "mean_token_accuracy": 0.9918572433292866,
+      "num_tokens": 31172652.0,
+      "step": 5860
+    },
+    {
+      "entropy": 0.019832546763041137,
+      "epoch": 0.6303517587939699,
+      "grad_norm": 2.25,
+      "learning_rate": 6.355517267676398e-06,
+      "loss": 0.022037458419799805,
+      "mean_token_accuracy": 0.9914021987468005,
+      "num_tokens": 31285622.0,
+      "step": 5880
+    },
+    {
+      "entropy": 0.021016045037094956,
+      "epoch": 0.6324958123953098,
+      "grad_norm": 1.9375,
+      "learning_rate": 6.290946003358175e-06,
+      "loss": 0.02272486835718155,
+      "mean_token_accuracy": 0.9905358534306288,
+      "num_tokens": 31391456.0,
+      "step": 5900
+    },
+    {
+      "entropy": 0.018993245228944033,
+      "epoch": 0.6346398659966499,
+      "grad_norm": 2.515625,
+      "learning_rate": 6.226553560637847e-06,
+      "loss": 0.02007617652416229,
+      "mean_token_accuracy": 0.9922155175358057,
+      "num_tokens": 31496166.0,
+      "step": 5920
+    },
+    {
+      "entropy": 0.0195752263855411,
+      "epoch": 0.63678391959799,
+      "grad_norm": 2.125,
+      "learning_rate": 6.162343044015784e-06,
+      "loss": 0.02081257700920105,
+      "mean_token_accuracy": 0.9921386137604713,
+      "num_tokens": 31605934.0,
+      "step": 5940
+    },
+    {
+      "entropy": 0.019383694695488885,
+      "epoch": 0.63892797319933,
+      "grad_norm": 2.046875,
+      "learning_rate": 6.098317549221302e-06,
+      "loss": 0.02021864652633667,
+      "mean_token_accuracy": 0.9916181784123183,
+      "num_tokens": 31712938.0,
+      "step": 5960
+    },
+    {
+      "entropy": 0.01924167173308433,
+      "epoch": 0.64107202680067,
+      "grad_norm": 2.046875,
+      "learning_rate": 6.034480163063401e-06,
+      "loss": 0.018102942407131194,
+      "mean_token_accuracy": 0.9921923886984587,
+      "num_tokens": 31813225.0,
+      "step": 5980
+    },
+    {
+      "entropy": 0.018873273960889492,
+      "epoch": 0.6432160804020101,
+      "grad_norm": 1.5234375,
+      "learning_rate": 5.970833963281968e-06,
+      "loss": 0.018628720939159394,
+      "mean_token_accuracy": 0.9928794108331204,
+      "num_tokens": 31928881.0,
+      "step": 6000
+    },
+    {
+      "epoch": 0.6432160804020101,
+      "eval_entropy": 0.01981416783077293,
+      "eval_loss": 0.019560638815164566,
+      "eval_mean_token_accuracy": 0.9923082728385926,
+      "eval_num_tokens": 31928881.0,
+      "eval_runtime": 20.2811,
+      "eval_samples_per_second": 36.98,
+      "eval_steps_per_second": 18.49,
+      "step": 6000
+    },
+    {
+      "entropy": 0.020261389288043573,
+      "epoch": 0.6453601340033501,
+      "grad_norm": 1.3125,
+      "learning_rate": 5.9073820183993635e-06,
+      "loss": 0.022888460755348207,
+      "mean_token_accuracy": 0.9925485417246819,
+      "num_tokens": 32038408.0,
+      "step": 6020
+    },
+    {
+      "entropy": 0.02017652262293268,
+      "epoch": 0.6475041876046901,
+      "grad_norm": 1.6328125,
+      "learning_rate": 5.8441273875725e-06,
+      "loss": 0.020387469232082366,
+      "mean_token_accuracy": 0.9918923225253821,
+      "num_tokens": 32145003.0,
+      "step": 6040
+    },
+    {
+      "entropy": 0.01868033191105951,
+      "epoch": 0.6496482412060302,
+      "grad_norm": 3.375,
+      "learning_rate": 5.781073120445353e-06,
+      "loss": 0.02243368774652481,
+      "mean_token_accuracy": 0.9915057402104139,
+      "num_tokens": 32251639.0,
+      "step": 6060
+    },
+    {
+      "entropy": 0.017508955025914475,
+      "epoch": 0.6517922948073702,
+      "grad_norm": 2.140625,
+      "learning_rate": 5.71822225700192e-06,
+      "loss": 0.017190653085708617,
+      "mean_token_accuracy": 0.9927597284317017,
+      "num_tokens": 32352371.0,
+      "step": 6080
+    },
+    {
+      "entropy": 0.019111776090062448,
+      "epoch": 0.6539363484087102,
+      "grad_norm": 1.1328125,
+      "learning_rate": 5.655577827419661e-06,
+      "loss": 0.02137412428855896,
+      "mean_token_accuracy": 0.9919875215739011,
+      "num_tokens": 32454655.0,
+      "step": 6100
+    },
+    {
+      "entropy": 0.023027155749787198,
+      "epoch": 0.6560804020100502,
+      "grad_norm": 2.015625,
+      "learning_rate": 5.5931428519234145e-06,
+      "loss": 0.02286381125450134,
+      "mean_token_accuracy": 0.9909462627023459,
+      "num_tokens": 32559559.0,
+      "step": 6120
+    },
+    {
+      "entropy": 0.02029506681608382,
+      "epoch": 0.6582244556113903,
+      "grad_norm": 3.21875,
+      "learning_rate": 5.530920340639776e-06,
+      "loss": 0.02172601819038391,
+      "mean_token_accuracy": 0.99092488437891,
+      "num_tokens": 32658365.0,
+      "step": 6140
+    },
+    {
+      "entropy": 0.021070733480109995,
+      "epoch": 0.6603685092127303,
+      "grad_norm": 3.625,
+      "learning_rate": 5.468913293451979e-06,
+      "loss": 0.021482250094413756,
+      "mean_token_accuracy": 0.9916323244571685,
+      "num_tokens": 32764039.0,
+      "step": 6160
+    },
+    {
+      "entropy": 0.018422811868731515,
+      "epoch": 0.6625125628140703,
+      "grad_norm": 2.6875,
+      "learning_rate": 5.407124699855259e-06,
+      "loss": 0.01822958290576935,
+      "mean_token_accuracy": 0.9923275396227836,
+      "num_tokens": 32877996.0,
+      "step": 6180
+    },
+    {
+      "entropy": 0.01931439164168296,
+      "epoch": 0.6646566164154104,
+      "grad_norm": 2.296875,
+      "learning_rate": 5.34555753881273e-06,
+      "loss": 0.021021611988544464,
+      "mean_token_accuracy": 0.9911792669445276,
+      "num_tokens": 32986191.0,
+      "step": 6200
+    },
+    {
+      "entropy": 0.019900657032940215,
+      "epoch": 0.6668006700167505,
+      "grad_norm": 1.734375,
+      "learning_rate": 5.284214778611758e-06,
+      "loss": 0.021909794211387633,
+      "mean_token_accuracy": 0.9917173352092504,
+      "num_tokens": 33098196.0,
+      "step": 6220
+    },
+    {
+      "entropy": 0.022017163155032902,
+      "epoch": 0.6689447236180904,
+      "grad_norm": 2.5625,
+      "learning_rate": 5.2230993767208524e-06,
+      "loss": 0.02269018590450287,
+      "mean_token_accuracy": 0.9912259686738253,
+      "num_tokens": 33211358.0,
+      "step": 6240
+    },
+    {
+      "entropy": 0.019263985499833325,
+      "epoch": 0.6710887772194305,
+      "grad_norm": 1.5625,
+      "learning_rate": 5.162214279647082e-06,
+      "loss": 0.023519453406333924,
+      "mean_token_accuracy": 0.9909868344664574,
+      "num_tokens": 33314953.0,
+      "step": 6260
+    },
+    {
+      "entropy": 0.021732687100166004,
+      "epoch": 0.6732328308207706,
+      "grad_norm": 2.265625,
+      "learning_rate": 5.1015624227940195e-06,
+      "loss": 0.021917779743671418,
+      "mean_token_accuracy": 0.9901779539883137,
+      "num_tokens": 33428051.0,
+      "step": 6280
+    },
+    {
+      "entropy": 0.019478338875705957,
+      "epoch": 0.6753768844221105,
+      "grad_norm": 2.296875,
+      "learning_rate": 5.041146730320215e-06,
+      "loss": 0.01874311566352844,
+      "mean_token_accuracy": 0.9923384670168162,
+      "num_tokens": 33537514.0,
+      "step": 6300
+    },
+    {
+      "entropy": 0.018508305449995533,
+      "epoch": 0.6775209380234506,
+      "grad_norm": 3.125,
+      "learning_rate": 4.980970114998216e-06,
+      "loss": 0.021730564534664154,
+      "mean_token_accuracy": 0.9915897861123085,
+      "num_tokens": 33639522.0,
+      "step": 6320
+    },
+    {
+      "entropy": 0.019665157535450818,
+      "epoch": 0.6796649916247907,
+      "grad_norm": 2.625,
+      "learning_rate": 4.921035478074142e-06,
+      "loss": 0.020458599925041197,
+      "mean_token_accuracy": 0.9918417375534773,
+      "num_tokens": 33743819.0,
+      "step": 6340
+    },
+    {
+      "entropy": 0.020174525574702784,
+      "epoch": 0.6818090452261306,
+      "grad_norm": 8.5,
+      "learning_rate": 4.861345709127803e-06,
+      "loss": 0.019577798247337342,
+      "mean_token_accuracy": 0.9921597026288509,
+      "num_tokens": 33847282.0,
+      "step": 6360
+    },
+    {
+      "entropy": 0.020318039059475268,
+      "epoch": 0.6839530988274707,
+      "grad_norm": 1.953125,
+      "learning_rate": 4.801903685933389e-06,
+      "loss": 0.020683969557285308,
+      "mean_token_accuracy": 0.9917214840650559,
+      "num_tokens": 33953961.0,
+      "step": 6380
+    },
+    {
+      "entropy": 0.01895209533279285,
+      "epoch": 0.6860971524288108,
+      "grad_norm": 2.046875,
+      "learning_rate": 4.742712274320725e-06,
+      "loss": 0.020713077485561372,
+      "mean_token_accuracy": 0.9920912489295006,
+      "num_tokens": 34054466.0,
+      "step": 6400
+    },
+    {
+      "entropy": 0.021405872826358065,
+      "epoch": 0.6882412060301507,
+      "grad_norm": 2.015625,
+      "learning_rate": 4.683774328037108e-06,
+      "loss": 0.024052520096302033,
+      "mean_token_accuracy": 0.9902271404862404,
+      "num_tokens": 34162751.0,
+      "step": 6420
+    },
+    {
+      "entropy": 0.02056694905677432,
+      "epoch": 0.6903852596314908,
+      "grad_norm": 2.109375,
+      "learning_rate": 4.625092688609711e-06,
+      "loss": 0.021541555225849152,
+      "mean_token_accuracy": 0.9906902097165584,
+      "num_tokens": 34271073.0,
+      "step": 6440
+    },
+    {
+      "entropy": 0.01796177522041944,
+      "epoch": 0.6925293132328308,
+      "grad_norm": 1.421875,
+      "learning_rate": 4.566670185208599e-06,
+      "loss": 0.017521923780441283,
+      "mean_token_accuracy": 0.9920081544667483,
+      "num_tokens": 34379697.0,
+      "step": 6460
+    },
+    {
+      "entropy": 0.01806882056660015,
+      "epoch": 0.6946733668341708,
+      "grad_norm": 1.6953125,
+      "learning_rate": 4.5085096345103205e-06,
+      "loss": 0.018725714087486266,
+      "mean_token_accuracy": 0.9868147686123848,
+      "num_tokens": 34483763.0,
+      "step": 6480
+    },
+    {
+      "entropy": 0.019470379299946217,
+      "epoch": 0.6968174204355109,
+      "grad_norm": 1.9140625,
+      "learning_rate": 4.450613840562113e-06,
+      "loss": 0.017752458155155183,
+      "mean_token_accuracy": 0.9928225323557853,
+      "num_tokens": 34588293.0,
+      "step": 6500
+    },
+    {
+      "epoch": 0.6968174204355109,
+      "eval_entropy": 0.019607524218813826,
+      "eval_loss": 0.019470153376460075,
+      "eval_mean_token_accuracy": 0.9920525077184041,
+      "eval_num_tokens": 34588293.0,
+      "eval_runtime": 20.2964,
+      "eval_samples_per_second": 36.952,
+      "eval_steps_per_second": 18.476,
+      "step": 6500
+    },
+    {
+      "entropy": 0.01848742664128622,
+      "epoch": 0.6989614740368509,
+      "grad_norm": 2.671875,
+      "learning_rate": 4.392985594646711e-06,
+      "loss": 0.02162926346063614,
+      "mean_token_accuracy": 0.9913080882281065,
+      "num_tokens": 34689227.0,
+      "step": 6520
+    },
+    {
+      "entropy": 0.019617287777646198,
+      "epoch": 0.701105527638191,
+      "grad_norm": 1.4765625,
+      "learning_rate": 4.335627675147774e-06,
+      "loss": 0.02028166800737381,
+      "mean_token_accuracy": 0.9930087491869927,
+      "num_tokens": 34791268.0,
+      "step": 6540
+    },
+    {
+      "entropy": 0.01906899958651138,
+      "epoch": 0.703249581239531,
+      "grad_norm": 2.234375,
+      "learning_rate": 4.278542847415936e-06,
+      "loss": 0.019539786875247954,
+      "mean_token_accuracy": 0.9920954536646605,
+      "num_tokens": 34894826.0,
+      "step": 6560
+    },
+    {
+      "entropy": 0.0211970092474985,
+      "epoch": 0.705393634840871,
+      "grad_norm": 1.5,
+      "learning_rate": 4.221733863635479e-06,
+      "loss": 0.023706386983394622,
+      "mean_token_accuracy": 0.990295572206378,
+      "num_tokens": 35009299.0,
+      "step": 6580
+    },
+    {
+      "entropy": 0.021638639217917442,
+      "epoch": 0.7075376884422111,
+      "grad_norm": 1.59375,
+      "learning_rate": 4.165203462691645e-06,
+      "loss": 0.021413305401802064,
+      "mean_token_accuracy": 0.9913092136383057,
+      "num_tokens": 35116618.0,
+      "step": 6600
+    },
+    {
+      "entropy": 0.02009705953132652,
+      "epoch": 0.7096817420435511,
+      "grad_norm": 1.9296875,
+      "learning_rate": 4.1089543700385885e-06,
+      "loss": 0.023159290850162505,
+      "mean_token_accuracy": 0.9899264223873615,
+      "num_tokens": 35214619.0,
+      "step": 6620
+    },
+    {
+      "entropy": 0.020903292103525928,
+      "epoch": 0.7118257956448911,
+      "grad_norm": 2.8125,
+      "learning_rate": 4.0529892975679775e-06,
+      "loss": 0.02029138207435608,
+      "mean_token_accuracy": 0.9917101729661226,
+      "num_tokens": 35331847.0,
+      "step": 6640
+    },
+    {
+      "entropy": 0.019355950780254715,
+      "epoch": 0.7139698492462312,
+      "grad_norm": 2.25,
+      "learning_rate": 3.997310943478246e-06,
+      "loss": 0.022747300565242767,
+      "mean_token_accuracy": 0.9911560833454132,
+      "num_tokens": 35435144.0,
+      "step": 6660
+    },
+    {
+      "entropy": 0.019431080123217724,
+      "epoch": 0.7161139028475711,
+      "grad_norm": 2.46875,
+      "learning_rate": 3.941921992144509e-06,
+      "loss": 0.022220295667648316,
+      "mean_token_accuracy": 0.9914675705134869,
+      "num_tokens": 35540146.0,
+      "step": 6680
+    },
+    {
+      "entropy": 0.021516031174610362,
+      "epoch": 0.7182579564489112,
+      "grad_norm": 1.0078125,
+      "learning_rate": 3.886825113989142e-06,
+      "loss": 0.020635062456130983,
+      "mean_token_accuracy": 0.9917585879564286,
+      "num_tokens": 35652696.0,
+      "step": 6700
+    },
+    {
+      "entropy": 0.019729190898897287,
+      "epoch": 0.7204020100502513,
+      "grad_norm": 2.09375,
+      "learning_rate": 3.832022965353032e-06,
+      "loss": 0.019225403666496277,
+      "mean_token_accuracy": 0.9942506503313779,
+      "num_tokens": 35757381.0,
+      "step": 6720
+    },
+    {
+      "entropy": 0.020393153377153795,
+      "epoch": 0.7225460636515912,
+      "grad_norm": 5.96875,
+      "learning_rate": 3.777518188367517e-06,
+      "loss": 0.02378985732793808,
+      "mean_token_accuracy": 0.9912960890680551,
+      "num_tokens": 35865336.0,
+      "step": 6740
+    },
+    {
+      "entropy": 0.019717699302668736,
+      "epoch": 0.7246901172529313,
+      "grad_norm": 1.8515625,
+      "learning_rate": 3.723313410826992e-06,
+      "loss": 0.020520621538162233,
+      "mean_token_accuracy": 0.9927110608667136,
+      "num_tokens": 35969262.0,
+      "step": 6760
+    },
+    {
+      "entropy": 0.020156670222718276,
+      "epoch": 0.7268341708542714,
+      "grad_norm": 2.953125,
+      "learning_rate": 3.6694112460622267e-06,
+      "loss": 0.022173659503459932,
+      "mean_token_accuracy": 0.9913631960749626,
+      "num_tokens": 36078203.0,
+      "step": 6780
+    },
+    {
+      "entropy": 0.02017759628906788,
+      "epoch": 0.7289782244556114,
+      "grad_norm": 1.6328125,
+      "learning_rate": 3.615814292814367e-06,
+      "loss": 0.02243554890155792,
+      "mean_token_accuracy": 0.9910614918917418,
+      "num_tokens": 36179468.0,
+      "step": 6800
+    },
+    {
+      "entropy": 0.019291495129346005,
+      "epoch": 0.7311222780569514,
+      "grad_norm": 2.03125,
+      "learning_rate": 3.562525135109649e-06,
+      "loss": 0.021489365398883818,
+      "mean_token_accuracy": 0.992356463521719,
+      "num_tokens": 36287569.0,
+      "step": 6820
+    },
+    {
+      "entropy": 0.017877134253399162,
+      "epoch": 0.7332663316582915,
+      "grad_norm": 2.8125,
+      "learning_rate": 3.509546342134796e-06,
+      "loss": 0.01892261952161789,
+      "mean_token_accuracy": 0.9919790286570788,
+      "num_tokens": 36403262.0,
+      "step": 6840
+    },
+    {
+      "entropy": 0.01958747923711144,
+      "epoch": 0.7354103852596315,
+      "grad_norm": 3.3125,
+      "learning_rate": 3.4568804681131964e-06,
+      "loss": 0.019973862171173095,
+      "mean_token_accuracy": 0.9915773533284664,
+      "num_tokens": 36508584.0,
+      "step": 6860
+    },
+    {
+      "entropy": 0.019319427383879883,
+      "epoch": 0.7375544388609715,
+      "grad_norm": 2.296875,
+      "learning_rate": 3.4045300521817105e-06,
+      "loss": 0.02136710286140442,
+      "mean_token_accuracy": 0.9920724961906672,
+      "num_tokens": 36610038.0,
+      "step": 6880
+    },
+    {
+      "entropy": 0.020166188537677952,
+      "epoch": 0.7396984924623116,
+      "grad_norm": 1.9453125,
+      "learning_rate": 3.3524976182682843e-06,
+      "loss": 0.018660889565944673,
+      "mean_token_accuracy": 0.9929303977638483,
+      "num_tokens": 36708085.0,
+      "step": 6900
+    },
+    {
+      "entropy": 0.018660859293049725,
+      "epoch": 0.7418425460636516,
+      "grad_norm": 2.40625,
+      "learning_rate": 3.3007856749702573e-06,
+      "loss": 0.01911092400550842,
+      "mean_token_accuracy": 0.9929109118878842,
+      "num_tokens": 36809564.0,
+      "step": 6920
+    },
+    {
+      "entropy": 0.020304086872602056,
+      "epoch": 0.7439865996649916,
+      "grad_norm": 1.734375,
+      "learning_rate": 3.2493967154334016e-06,
+      "loss": 0.019811296463012697,
+      "mean_token_accuracy": 0.9924409098923206,
+      "num_tokens": 36912901.0,
+      "step": 6940
+    },
+    {
+      "entropy": 0.017767412251100723,
+      "epoch": 0.7461306532663317,
+      "grad_norm": 1.515625,
+      "learning_rate": 3.198333217231757e-06,
+      "loss": 0.016918018460273743,
+      "mean_token_accuracy": 0.9936570934951305,
+      "num_tokens": 37028495.0,
+      "step": 6960
+    },
+    {
+      "entropy": 0.01919960001034724,
+      "epoch": 0.7482747068676717,
+      "grad_norm": 1.78125,
+      "learning_rate": 3.1475976422481526e-06,
+      "loss": 0.018871791660785675,
+      "mean_token_accuracy": 0.9923384845256805,
+      "num_tokens": 37128754.0,
+      "step": 6980
+    },
+    {
+      "entropy": 0.01957385808059371,
+      "epoch": 0.7504187604690117,
+      "grad_norm": 2.515625,
+      "learning_rate": 3.0971924365555116e-06,
+      "loss": 0.018768243491649628,
+      "mean_token_accuracy": 0.9924201011657715,
+      "num_tokens": 37230511.0,
+      "step": 7000
+    },
+    {
+      "epoch": 0.7504187604690117,
+      "eval_entropy": 0.019512116880068788,
+      "eval_loss": 0.019492411985993385,
+      "eval_mean_token_accuracy": 0.9919790083567301,
+      "eval_num_tokens": 37230511.0,
+      "eval_runtime": 20.2777,
+      "eval_samples_per_second": 36.986,
+      "eval_steps_per_second": 18.493,
+      "step": 7000
+    },
+    {
+      "entropy": 0.018825849278618988,
+      "epoch": 0.7525628140703517,
+      "grad_norm": 2.265625,
+      "learning_rate": 3.0471200302989535e-06,
+      "loss": 0.020853203535079957,
+      "mean_token_accuracy": 0.9906294353306293,
+      "num_tokens": 37332200.0,
+      "step": 7020
+    },
+    {
+      "entropy": 0.019942389463130893,
+      "epoch": 0.7547068676716918,
+      "grad_norm": 2.203125,
+      "learning_rate": 2.9973828375785864e-06,
+      "loss": 0.018565280735492705,
+      "mean_token_accuracy": 0.9909187290817499,
+      "num_tokens": 37433333.0,
+      "step": 7040
+    },
+    {
+      "entropy": 0.020680780549423617,
+      "epoch": 0.7568509212730318,
+      "grad_norm": 1.3984375,
+      "learning_rate": 2.947983256333166e-06,
+      "loss": 0.022726891934871672,
+      "mean_token_accuracy": 0.9920847479254007,
+      "num_tokens": 37535155.0,
+      "step": 7060
+    },
+    {
+      "entropy": 0.01799933227546262,
+      "epoch": 0.7589949748743718,
+      "grad_norm": 2.375,
+      "learning_rate": 2.8989236682244372e-06,
+      "loss": 0.019517210125923157,
+      "mean_token_accuracy": 0.9922154374420643,
+      "num_tokens": 37642429.0,
+      "step": 7080
+    },
+    {
+      "entropy": 0.018576059823135437,
+      "epoch": 0.7611390284757119,
+      "grad_norm": 1.8671875,
+      "learning_rate": 2.8502064385223437e-06,
+      "loss": 0.02000274509191513,
+      "mean_token_accuracy": 0.9927195519208908,
+      "num_tokens": 37743382.0,
+      "step": 7100
+    },
+    {
+      "entropy": 0.01817750639111182,
+      "epoch": 0.763283082077052,
+      "grad_norm": 0.76953125,
+      "learning_rate": 2.8018339159909937e-06,
+      "loss": 0.018051585555076598,
+      "mean_token_accuracy": 0.9927114237099885,
+      "num_tokens": 37849802.0,
+      "step": 7120
+    },
+    {
+      "entropy": 0.01861468399677051,
+      "epoch": 0.7654271356783919,
+      "grad_norm": 2.453125,
+      "learning_rate": 2.7538084327753823e-06,
+      "loss": 0.020044369995594023,
+      "mean_token_accuracy": 0.9925104282796383,
+      "num_tokens": 37949510.0,
+      "step": 7140
+    },
+    {
+      "entropy": 0.01957492132596599,
+      "epoch": 0.767571189279732,
+      "grad_norm": 1.6640625,
+      "learning_rate": 2.706132304289012e-06,
+      "loss": 0.02030598223209381,
+      "mean_token_accuracy": 0.9924644228070975,
+      "num_tokens": 38056600.0,
+      "step": 7160
+    },
+    {
+      "entropy": 0.019891365615512767,
+      "epoch": 0.7697152428810721,
+      "grad_norm": 1.3046875,
+      "learning_rate": 2.6588078291022045e-06,
+      "loss": 0.020326973497867586,
+      "mean_token_accuracy": 0.9918508928269147,
+      "num_tokens": 38168190.0,
+      "step": 7180
+    },
+    {
+      "entropy": 0.01919556673688021,
+      "epoch": 0.771859296482412,
+      "grad_norm": 2.796875,
+      "learning_rate": 2.611837288831319e-06,
+      "loss": 0.02155778855085373,
+      "mean_token_accuracy": 0.9919852402061224,
+      "num_tokens": 38276246.0,
+      "step": 7200
+    },
+    {
+      "entropy": 0.019413144076725074,
+      "epoch": 0.7740033500837521,
+      "grad_norm": 5.8125,
+      "learning_rate": 2.565222948028746e-06,
+      "loss": 0.0233286514878273,
+      "mean_token_accuracy": 0.9901302725076675,
+      "num_tokens": 38380173.0,
+      "step": 7220
+    },
+    {
+      "entropy": 0.019048391740625448,
+      "epoch": 0.7761474036850922,
+      "grad_norm": 2.484375,
+      "learning_rate": 2.518967054073702e-06,
+      "loss": 0.01727575808763504,
+      "mean_token_accuracy": 0.9933784183114767,
+      "num_tokens": 38485326.0,
+      "step": 7240
+    },
+    {
+      "entropy": 0.018992443338993324,
+      "epoch": 0.7782914572864321,
+      "grad_norm": 1.765625,
+      "learning_rate": 2.473071837063921e-06,
+      "loss": 0.020196519792079926,
+      "mean_token_accuracy": 0.9929014749825,
+      "num_tokens": 38593033.0,
+      "step": 7260
+    },
+    {
+      "entropy": 0.01934261324011004,
+      "epoch": 0.7804355108877722,
+      "grad_norm": 2.078125,
+      "learning_rate": 2.427539509708092e-06,
+      "loss": 0.021069280803203583,
+      "mean_token_accuracy": 0.9920600451529026,
+      "num_tokens": 38695209.0,
+      "step": 7280
+    },
+    {
+      "entropy": 0.01937621497809232,
+      "epoch": 0.7825795644891123,
+      "grad_norm": 1.96875,
+      "learning_rate": 2.382372267219212e-06,
+      "loss": 0.02313290983438492,
+      "mean_token_accuracy": 0.9915325742214918,
+      "num_tokens": 38804750.0,
+      "step": 7300
+    },
+    {
+      "entropy": 0.02088976109387204,
+      "epoch": 0.7847236180904522,
+      "grad_norm": 3.390625,
+      "learning_rate": 2.3375722872087435e-06,
+      "loss": 0.02172703593969345,
+      "mean_token_accuracy": 0.9911188166588545,
+      "num_tokens": 38909561.0,
+      "step": 7320
+    },
+    {
+      "entropy": 0.019513376194367993,
+      "epoch": 0.7868676716917923,
+      "grad_norm": 1.6171875,
+      "learning_rate": 2.293141729581606e-06,
+      "loss": 0.020942328870296477,
+      "mean_token_accuracy": 0.9914060894399881,
+      "num_tokens": 39024348.0,
+      "step": 7340
+    },
+    {
+      "entropy": 0.019013147754321835,
+      "epoch": 0.7890117252931323,
+      "grad_norm": 3.03125,
+      "learning_rate": 2.24908273643208e-06,
+      "loss": 0.01895921677350998,
+      "mean_token_accuracy": 0.9920435093343258,
+      "num_tokens": 39123358.0,
+      "step": 7360
+    },
+    {
+      "entropy": 0.01850612224473025,
+      "epoch": 0.7911557788944724,
+      "grad_norm": 2.0625,
+      "learning_rate": 2.20539743194049e-06,
+      "loss": 0.020412975549697877,
+      "mean_token_accuracy": 0.992505956813693,
+      "num_tokens": 39222173.0,
+      "step": 7380
+    },
+    {
+      "entropy": 0.018990432533109926,
+      "epoch": 0.7932998324958124,
+      "grad_norm": 2.515625,
+      "learning_rate": 2.162087922270827e-06,
+      "loss": 0.019987034797668456,
+      "mean_token_accuracy": 0.9921554632484912,
+      "num_tokens": 39324555.0,
+      "step": 7400
+    },
+    {
+      "entropy": 0.020331120665650813,
+      "epoch": 0.7954438860971524,
+      "grad_norm": 2.59375,
+      "learning_rate": 2.119156295469188e-06,
+      "loss": 0.020988868176937105,
+      "mean_token_accuracy": 0.992167953774333,
+      "num_tokens": 39436930.0,
+      "step": 7420
+    },
+    {
+      "entropy": 0.01908351750444126,
+      "epoch": 0.7975879396984925,
+      "grad_norm": 2.140625,
+      "learning_rate": 2.0766046213631085e-06,
+      "loss": 0.02441962659358978,
+      "mean_token_accuracy": 0.9905336666852236,
+      "num_tokens": 39539754.0,
+      "step": 7440
+    },
+    {
+      "entropy": 0.01844053882264234,
+      "epoch": 0.7997319932998325,
+      "grad_norm": 1.4609375,
+      "learning_rate": 2.034434951461787e-06,
+      "loss": 0.021246211230754854,
+      "mean_token_accuracy": 0.9916405260562897,
+      "num_tokens": 39642853.0,
+      "step": 7460
+    },
+    {
+      "entropy": 0.019568845091521327,
+      "epoch": 0.8018760469011725,
+      "grad_norm": 2.234375,
+      "learning_rate": 1.992649318857145e-06,
+      "loss": 0.02266250252723694,
+      "mean_token_accuracy": 0.9901510171592236,
+      "num_tokens": 39753649.0,
+      "step": 7480
+    },
+    {
+      "entropy": 0.019649790661560473,
+      "epoch": 0.8040201005025126,
+      "grad_norm": 1.4296875,
+      "learning_rate": 1.951249738125841e-06,
+      "loss": 0.02106764316558838,
+      "mean_token_accuracy": 0.992297976091504,
+      "num_tokens": 39867091.0,
+      "step": 7500
+    },
+    {
+      "epoch": 0.8040201005025126,
+      "eval_entropy": 0.01951531241268579,
+      "eval_loss": 0.019423136487603188,
+      "eval_mean_token_accuracy": 0.9921902815500895,
+      "eval_num_tokens": 39867091.0,
+      "eval_runtime": 20.2653,
+      "eval_samples_per_second": 37.009,
+      "eval_steps_per_second": 18.504,
+      "step": 7500
+    },
+    {
+      "entropy": 0.01873672366637038,
+      "epoch": 0.8061641541038526,
+      "grad_norm": 2.515625,
+      "learning_rate": 1.9102382052321247e-06,
+      "loss": 0.02217361479997635,
+      "mean_token_accuracy": 0.991604308411479,
+      "num_tokens": 39975152.0,
+      "step": 7520
+    },
+    {
+      "entropy": 0.018056424487326694,
+      "epoch": 0.8083082077051926,
+      "grad_norm": 2.1875,
+      "learning_rate": 1.8696166974316098e-06,
+      "loss": 0.018713992834091187,
+      "mean_token_accuracy": 0.9923322558403015,
+      "num_tokens": 40073850.0,
+      "step": 7540
+    },
+    {
+      "entropy": 0.019656274415865482,
+      "epoch": 0.8104522613065327,
+      "grad_norm": 1.75,
+      "learning_rate": 1.829387173175957e-06,
+      "loss": 0.02295905351638794,
+      "mean_token_accuracy": 0.9911978255957365,
+      "num_tokens": 40174029.0,
+      "step": 7560
+    },
+    {
+      "entropy": 0.018212407533337683,
+      "epoch": 0.8125963149078727,
+      "grad_norm": 2.203125,
+      "learning_rate": 1.7895515720184298e-06,
+      "loss": 0.01762194335460663,
+      "mean_token_accuracy": 0.993547822907567,
+      "num_tokens": 40277338.0,
+      "step": 7580
+    },
+    {
+      "entropy": 0.019111268576693874,
+      "epoch": 0.8147403685092127,
+      "grad_norm": 3.078125,
+      "learning_rate": 1.7501118145204078e-06,
+      "loss": 0.02163155972957611,
+      "mean_token_accuracy": 0.9915014613419771,
+      "num_tokens": 40384919.0,
+      "step": 7600
+    },
+    {
+      "entropy": 0.020142577382034687,
+      "epoch": 0.8168844221105528,
+      "grad_norm": 2.1875,
+      "learning_rate": 1.71106980215878e-06,
+      "loss": 0.01972368061542511,
+      "mean_token_accuracy": 0.9922790337353945,
+      "num_tokens": 40491708.0,
+      "step": 7620
+    },
+    {
+      "entropy": 0.019609718415108546,
+      "epoch": 0.8190284757118927,
+      "grad_norm": 6.25,
+      "learning_rate": 1.6724274172342725e-06,
+      "loss": 0.023548129200935363,
+      "mean_token_accuracy": 0.991268851980567,
+      "num_tokens": 40604052.0,
+      "step": 7640
+    },
+    {
+      "entropy": 0.019134915648101013,
+      "epoch": 0.8211725293132328,
+      "grad_norm": 1.75,
+      "learning_rate": 1.6341865227807075e-06,
+      "loss": 0.019575177133083342,
+      "mean_token_accuracy": 0.992620923742652,
+      "num_tokens": 40712926.0,
+      "step": 7660
+    },
+    {
+      "entropy": 0.018662033730379335,
+      "epoch": 0.8233165829145729,
+      "grad_norm": 2.265625,
+      "learning_rate": 1.5963489624751595e-06,
+      "loss": 0.019268378615379333,
+      "mean_token_accuracy": 0.992533054947853,
+      "num_tokens": 40818181.0,
+      "step": 7680
+    },
+    {
+      "entropy": 0.019738298160768865,
+      "epoch": 0.825460636515913,
+      "grad_norm": 1.9921875,
+      "learning_rate": 1.5589165605490918e-06,
+      "loss": 0.018965110182762146,
+      "mean_token_accuracy": 0.9918511062860489,
+      "num_tokens": 40925382.0,
+      "step": 7700
+    },
+    {
+      "entropy": 0.01947517461164807,
+      "epoch": 0.8276046901172529,
+      "grad_norm": 2.046875,
+      "learning_rate": 1.5218911217003985e-06,
+      "loss": 0.02183605879545212,
+      "mean_token_accuracy": 0.9911946535110474,
+      "num_tokens": 41044722.0,
+      "step": 7720
+    },
+    {
+      "entropy": 0.019855259767928145,
+      "epoch": 0.829748743718593,
+      "grad_norm": 1.90625,
+      "learning_rate": 1.4852744310063893e-06,
+      "loss": 0.020095731317996978,
+      "mean_token_accuracy": 0.9914216987788678,
+      "num_tokens": 41163232.0,
+      "step": 7740
+    },
+    {
+      "entropy": 0.01931553952094873,
+      "epoch": 0.831892797319933,
+      "grad_norm": 2.703125,
+      "learning_rate": 1.4490682538377344e-06,
+      "loss": 0.02068305015563965,
+      "mean_token_accuracy": 0.9912440560758113,
+      "num_tokens": 41275818.0,
+      "step": 7760
+    },
+    {
+      "entropy": 0.01949581656031114,
+      "epoch": 0.834036850921273,
+      "grad_norm": 2.640625,
+      "learning_rate": 1.4132743357733502e-06,
+      "loss": 0.019521139562129974,
+      "mean_token_accuracy": 0.9927113294601441,
+      "num_tokens": 41378446.0,
+      "step": 7780
+    },
+    {
+      "entropy": 0.019279925492878647,
+      "epoch": 0.8361809045226131,
+      "grad_norm": 1.71875,
+      "learning_rate": 1.3778944025162399e-06,
+      "loss": 0.020213739573955537,
+      "mean_token_accuracy": 0.9918138355016708,
+      "num_tokens": 41480586.0,
+      "step": 7800
+    },
+    {
+      "entropy": 0.01848278012148512,
+      "epoch": 0.8383249581239531,
+      "grad_norm": 1.8046875,
+      "learning_rate": 1.3429301598102973e-06,
+      "loss": 0.021417385339736937,
+      "mean_token_accuracy": 0.992524030432105,
+      "num_tokens": 41583797.0,
+      "step": 7820
+    },
+    {
+      "entropy": 0.019653417706990695,
+      "epoch": 0.8404690117252931,
+      "grad_norm": 1.4375,
+      "learning_rate": 1.3083832933580608e-06,
+      "loss": 0.019776181876659395,
+      "mean_token_accuracy": 0.9927833370864392,
+      "num_tokens": 41693464.0,
+      "step": 7840
+    },
+    {
+      "entropy": 0.019763251290942207,
+      "epoch": 0.8426130653266332,
+      "grad_norm": 3.109375,
+      "learning_rate": 1.2742554687394537e-06,
+      "loss": 0.02297937124967575,
+      "mean_token_accuracy": 0.992657694220543,
+      "num_tokens": 41790157.0,
+      "step": 7860
+    },
+    {
+      "entropy": 0.019936064798548613,
+      "epoch": 0.8447571189279732,
+      "grad_norm": 3.609375,
+      "learning_rate": 1.2405483313314738e-06,
+      "loss": 0.02262817770242691,
+      "mean_token_accuracy": 0.9905056063085794,
+      "num_tokens": 41893582.0,
+      "step": 7880
+    },
+    {
+      "entropy": 0.01934517203821997,
+      "epoch": 0.8469011725293132,
+      "grad_norm": 2.109375,
+      "learning_rate": 1.2072635062288673e-06,
+      "loss": 0.020036940276622773,
+      "mean_token_accuracy": 0.9923514679074288,
+      "num_tokens": 42004717.0,
+      "step": 7900
+    },
+    {
+      "entropy": 0.01836295023481398,
+      "epoch": 0.8490452261306533,
+      "grad_norm": 3.1875,
+      "learning_rate": 1.174402598165787e-06,
+      "loss": 0.017996013164520264,
+      "mean_token_accuracy": 0.9931914895772934,
+      "num_tokens": 42114536.0,
+      "step": 7920
+    },
+    {
+      "entropy": 0.020249022626080658,
+      "epoch": 0.8511892797319933,
+      "grad_norm": 2.5,
+      "learning_rate": 1.1419671914384134e-06,
+      "loss": 0.02248285710811615,
+      "mean_token_accuracy": 0.991206606850028,
+      "num_tokens": 42222812.0,
+      "step": 7940
+    },
+    {
+      "entropy": 0.02073315970085332,
+      "epoch": 0.8533333333333334,
+      "grad_norm": 3.0,
+      "learning_rate": 1.1099588498285806e-06,
+      "loss": 0.02276255041360855,
+      "mean_token_accuracy": 0.9906704191118478,
+      "num_tokens": 42337872.0,
+      "step": 7960
+    },
+    {
+      "entropy": 0.019947545273566902,
+      "epoch": 0.8554773869346733,
+      "grad_norm": 1.5390625,
+      "learning_rate": 1.0783791165283798e-06,
+      "loss": 0.0205001100897789,
+      "mean_token_accuracy": 0.9920544352382421,
+      "num_tokens": 42443680.0,
+      "step": 7980
+    },
+    {
+      "entropy": 0.017547584273529537,
+      "epoch": 0.8576214405360134,
+      "grad_norm": 3.328125,
+      "learning_rate": 1.0472295140657584e-06,
+      "loss": 0.01985829621553421,
+      "mean_token_accuracy": 0.9921385683119297,
+      "num_tokens": 42548237.0,
+      "step": 8000
+    },
+    {
+      "epoch": 0.8576214405360134,
+      "eval_entropy": 0.019516329319584958,
+      "eval_loss": 0.01938561163842678,
+      "eval_mean_token_accuracy": 0.9923241523106893,
+      "eval_num_tokens": 42548237.0,
+      "eval_runtime": 20.235,
+      "eval_samples_per_second": 37.064,
+      "eval_steps_per_second": 18.532,
+      "step": 8000
+    },
+    {
+      "entropy": 0.018809675735496967,
+      "epoch": 0.8597654941373535,
+      "grad_norm": 1.5859375,
+      "learning_rate": 1.0165115442311168e-06,
+      "loss": 0.02014607787132263,
+      "mean_token_accuracy": 0.9919284801930189,
+      "num_tokens": 42659432.0,
+      "step": 8020
+    },
+    {
+      "entropy": 0.01733278074198097,
+      "epoch": 0.8619095477386934,
+      "grad_norm": 1.921875,
+      "learning_rate": 9.862266880049033e-07,
+      "loss": 0.018754176795482635,
+      "mean_token_accuracy": 0.9923460192978382,
+      "num_tokens": 42759466.0,
+      "step": 8040
+    },
+    {
+      "entropy": 0.01832941572265554,
+      "epoch": 0.8640536013400335,
+      "grad_norm": 2.078125,
+      "learning_rate": 9.563764054862112e-07,
+      "loss": 0.020297129452228547,
+      "mean_token_accuracy": 0.9928287036716938,
+      "num_tokens": 42868393.0,
+      "step": 8060
+    },
+    {
+      "entropy": 0.020013672889513147,
+      "epoch": 0.8661976549413736,
+      "grad_norm": 2.296875,
+      "learning_rate": 9.26962135822389e-07,
+      "loss": 0.021075569093227386,
+      "mean_token_accuracy": 0.9927308529615402,
+      "num_tokens": 42978572.0,
+      "step": 8080
+    },
+    {
+      "entropy": 0.017360182798279312,
+      "epoch": 0.8683417085427135,
+      "grad_norm": 1.5625,
+      "learning_rate": 8.979852971396508e-07,
+      "loss": 0.018198615312576293,
+      "mean_token_accuracy": 0.993445823341608,
+      "num_tokens": 43086831.0,
+      "step": 8100
+    },
+    {
+      "entropy": 0.02018861417880089,
+      "epoch": 0.8704857621440536,
+      "grad_norm": 2.25,
+      "learning_rate": 8.69447286474705e-07,
+      "loss": 0.02233450710773468,
+      "mean_token_accuracy": 0.9902903474867344,
+      "num_tokens": 43195060.0,
+      "step": 8120
+    },
+    {
+      "entropy": 0.019654828006423486,
+      "epoch": 0.8726298157453937,
+      "grad_norm": 2.515625,
+      "learning_rate": 8.413494797074095e-07,
+      "loss": 0.01920991241931915,
+      "mean_token_accuracy": 0.9916286468505859,
+      "num_tokens": 43306019.0,
+      "step": 8140
+    },
+    {
+      "entropy": 0.02208902873887837,
+      "epoch": 0.8747738693467336,
+      "grad_norm": 2.140625,
+      "learning_rate": 8.136932314944179e-07,
+      "loss": 0.022708958387374877,
+      "mean_token_accuracy": 0.9901477582752705,
+      "num_tokens": 43416895.0,
+      "step": 8160
+    },
+    {
+      "entropy": 0.01981349259772287,
+      "epoch": 0.8769179229480737,
+      "grad_norm": 2.71875,
+      "learning_rate": 7.864798752038949e-07,
+      "loss": 0.01928192377090454,
+      "mean_token_accuracy": 0.9919186655431986,
+      "num_tokens": 43516044.0,
+      "step": 8180
+    },
+    {
+      "entropy": 0.01925912519070607,
+      "epoch": 0.8790619765494138,
+      "grad_norm": 1.734375,
+      "learning_rate": 7.597107228512101e-07,
+      "loss": 0.02041231542825699,
+      "mean_token_accuracy": 0.9925541203469038,
+      "num_tokens": 43622616.0,
+      "step": 8200
+    },
+    {
+      "entropy": 0.01794401331856079,
+      "epoch": 0.8812060301507537,
+      "grad_norm": 1.40625,
+      "learning_rate": 7.333870650356911e-07,
+      "loss": 0.01903040409088135,
+      "mean_token_accuracy": 0.9926789086312056,
+      "num_tokens": 43738048.0,
+      "step": 8220
+    },
+    {
+      "entropy": 0.020185897200371984,
+      "epoch": 0.8833500837520938,
+      "grad_norm": 3.078125,
+      "learning_rate": 7.075101708784038e-07,
+      "loss": 0.02289024293422699,
+      "mean_token_accuracy": 0.9907027028501034,
+      "num_tokens": 43849936.0,
+      "step": 8240
+    },
+    {
+      "entropy": 0.020925003469278634,
+      "epoch": 0.8854941373534339,
+      "grad_norm": 2.96875,
+      "learning_rate": 6.820812879609551e-07,
+      "loss": 0.02114592641592026,
+      "mean_token_accuracy": 0.9914819072932005,
+      "num_tokens": 43953914.0,
+      "step": 8260
+    },
+    {
+      "entropy": 0.019342410219724115,
+      "epoch": 0.8876381909547739,
+      "grad_norm": 3.609375,
+      "learning_rate": 6.571016422653609e-07,
+      "loss": 0.020218712091445924,
+      "mean_token_accuracy": 0.9917961444705725,
+      "num_tokens": 44060574.0,
+      "step": 8280
+    },
+    {
+      "entropy": 0.018030824407537692,
+      "epoch": 0.8897822445561139,
+      "grad_norm": 1.171875,
+      "learning_rate": 6.325724381149201e-07,
+      "loss": 0.01951337456703186,
+      "mean_token_accuracy": 0.9863187871873379,
+      "num_tokens": 44169435.0,
+      "step": 8300
+    },
+    {
+      "entropy": 0.01994664284213741,
+      "epoch": 0.891926298157454,
+      "grad_norm": 2.921875,
+      "learning_rate": 6.084948581161643e-07,
+      "loss": 0.019201353192329407,
+      "mean_token_accuracy": 0.9925763100385666,
+      "num_tokens": 44276020.0,
+      "step": 8320
+    },
+    {
+      "entropy": 0.02081800427977214,
+      "epoch": 0.894070351758794,
+      "grad_norm": 1.6953125,
+      "learning_rate": 5.848700631018411e-07,
+      "loss": 0.020571966469287873,
+      "mean_token_accuracy": 0.9917529292404652,
+      "num_tokens": 44386885.0,
+      "step": 8340
+    },
+    {
+      "entropy": 0.01815503019108746,
+      "epoch": 0.896214405360134,
+      "grad_norm": 2.203125,
+      "learning_rate": 5.616991920749349e-07,
+      "loss": 0.020456781983375548,
+      "mean_token_accuracy": 0.9919084422290325,
+      "num_tokens": 44493234.0,
+      "step": 8360
+    },
+    {
+      "entropy": 0.018217097373872095,
+      "epoch": 0.898358458961474,
+      "grad_norm": 1.375,
+      "learning_rate": 5.389833621537755e-07,
+      "loss": 0.01739330738782883,
+      "mean_token_accuracy": 0.9934280533343554,
+      "num_tokens": 44596525.0,
+      "step": 8380
+    },
+    {
+      "entropy": 0.01944470095006636,
+      "epoch": 0.9005025125628141,
+      "grad_norm": 1.875,
+      "learning_rate": 5.167236685181532e-07,
+      "loss": 0.01959017515182495,
+      "mean_token_accuracy": 0.9926484901458025,
+      "num_tokens": 44704672.0,
+      "step": 8400
+    },
+    {
+      "entropy": 0.01922774040344848,
+      "epoch": 0.9026465661641541,
+      "grad_norm": 1.8671875,
+      "learning_rate": 4.949211843565449e-07,
+      "loss": 0.018985719978809358,
+      "mean_token_accuracy": 0.9930604692548514,
+      "num_tokens": 44802160.0,
+      "step": 8420
+    },
+    {
+      "entropy": 0.019086347535608183,
+      "epoch": 0.9047906197654941,
+      "grad_norm": 2.625,
+      "learning_rate": 4.7357696081435036e-07,
+      "loss": 0.021893636882305147,
+      "mean_token_accuracy": 0.9909486912190915,
+      "num_tokens": 44907874.0,
+      "step": 8440
+    },
+    {
+      "entropy": 0.01974412333381679,
+      "epoch": 0.9069346733668342,
+      "grad_norm": 3.1875,
+      "learning_rate": 4.5269202694322334e-07,
+      "loss": 0.0225388765335083,
+      "mean_token_accuracy": 0.9913734450936318,
+      "num_tokens": 45011015.0,
+      "step": 8460
+    },
+    {
+      "entropy": 0.0196786193088883,
+      "epoch": 0.9090787269681742,
+      "grad_norm": 3.484375,
+      "learning_rate": 4.3226738965146733e-07,
+      "loss": 0.019231879711151124,
+      "mean_token_accuracy": 0.9921259760856629,
+      "num_tokens": 45122211.0,
+      "step": 8480
+    },
+    {
+      "entropy": 0.017996370965624918,
+      "epoch": 0.9112227805695142,
+      "grad_norm": 1.3125,
+      "learning_rate": 4.1230403365547065e-07,
+      "loss": 0.018080990016460418,
+      "mean_token_accuracy": 0.9934276740998029,
+      "num_tokens": 45229362.0,
+      "step": 8500
+    },
+    {
+      "epoch": 0.9112227805695142,
+      "eval_entropy": 0.019543887719133637,
+      "eval_loss": 0.019348351284861565,
+      "eval_mean_token_accuracy": 0.9922665920257568,
+      "eval_num_tokens": 45229362.0,
+      "eval_runtime": 20.2762,
+      "eval_samples_per_second": 36.989,
+      "eval_steps_per_second": 18.495,
+      "step": 8500
+    },
+    {
+      "entropy": 0.021203386071078967,
+      "epoch": 0.9133668341708543,
+      "grad_norm": 1.609375,
+      "learning_rate": 3.9280292143225264e-07,
+      "loss": 0.022594137489795683,
+      "mean_token_accuracy": 0.991113456711173,
+      "num_tokens": 45338204.0,
+      "step": 8520
+    },
+    {
+      "entropy": 0.018270298717538935,
+      "epoch": 0.9155108877721944,
+      "grad_norm": 2.03125,
+      "learning_rate": 3.737649931730458e-07,
+      "loss": 0.019554194808006287,
+      "mean_token_accuracy": 0.9928128100931645,
+      "num_tokens": 45451611.0,
+      "step": 8540
+    },
+    {
+      "entropy": 0.01802048790036679,
+      "epoch": 0.9176549413735343,
+      "grad_norm": 1.8984375,
+      "learning_rate": 3.5519116673796707e-07,
+      "loss": 0.01832857131958008,
+      "mean_token_accuracy": 0.9920665070414543,
+      "num_tokens": 45552742.0,
+      "step": 8560
+    },
+    {
+      "entropy": 0.019545484634431887,
+      "epoch": 0.9197989949748744,
+      "grad_norm": 2.078125,
+      "learning_rate": 3.3708233761177935e-07,
+      "loss": 0.019848516583442687,
+      "mean_token_accuracy": 0.9930632192641496,
+      "num_tokens": 45654718.0,
+      "step": 8580
+    },
+    {
+      "entropy": 0.019058815206676627,
+      "epoch": 0.9219430485762145,
+      "grad_norm": 2.21875,
+      "learning_rate": 3.1943937886070197e-07,
+      "loss": 0.019996930658817292,
+      "mean_token_accuracy": 0.9920886129140853,
+      "num_tokens": 45764177.0,
+      "step": 8600
+    },
+    {
+      "entropy": 0.019873253857394958,
+      "epoch": 0.9240871021775544,
+      "grad_norm": 1.7265625,
+      "learning_rate": 3.0226314109032984e-07,
+      "loss": 0.020348772406578064,
+      "mean_token_accuracy": 0.9924048013985157,
+      "num_tokens": 45869501.0,
+      "step": 8620
+    },
+    {
+      "entropy": 0.01988984109498233,
+      "epoch": 0.9262311557788945,
+      "grad_norm": 2.09375,
+      "learning_rate": 2.8555445240461834e-07,
+      "loss": 0.01935010850429535,
+      "mean_token_accuracy": 0.9919997557997704,
+      "num_tokens": 45970153.0,
+      "step": 8640
+    },
+    {
+      "entropy": 0.01982094865072668,
+      "epoch": 0.9283752093802345,
+      "grad_norm": 2.09375,
+      "learning_rate": 2.6931411836595643e-07,
+      "loss": 0.02122329771518707,
+      "mean_token_accuracy": 0.9918153654783964,
+      "num_tokens": 46071225.0,
+      "step": 8660
+    },
+    {
+      "entropy": 0.019260422117486086,
+      "epoch": 0.9305192629815745,
+      "grad_norm": 2.359375,
+      "learning_rate": 2.535429219563401e-07,
+      "loss": 0.0204538568854332,
+      "mean_token_accuracy": 0.9913510870188474,
+      "num_tokens": 46180945.0,
+      "step": 8680
+    },
+    {
+      "entropy": 0.019332679143190035,
+      "epoch": 0.9326633165829146,
+      "grad_norm": 1.6796875,
+      "learning_rate": 2.3824162353960457e-07,
+      "loss": 0.02263512313365936,
+      "mean_token_accuracy": 0.9922845117747784,
+      "num_tokens": 46287275.0,
+      "step": 8700
+    },
+    {
+      "entropy": 0.018464213076777015,
+      "epoch": 0.9348073701842546,
+      "grad_norm": 2.046875,
+      "learning_rate": 2.2341096082478698e-07,
+      "loss": 0.019891925156116486,
+      "mean_token_accuracy": 0.9930620789527893,
+      "num_tokens": 46396075.0,
+      "step": 8720
+    },
+    {
+      "entropy": 0.020736305239142893,
+      "epoch": 0.9369514237855946,
+      "grad_norm": 1.796875,
+      "learning_rate": 2.0905164883054053e-07,
+      "loss": 0.0205291748046875,
+      "mean_token_accuracy": 0.9927501816302537,
+      "num_tokens": 46503373.0,
+      "step": 8740
+    },
+    {
+      "entropy": 0.01994141639102054,
+      "epoch": 0.9390954773869347,
+      "grad_norm": 1.8671875,
+      "learning_rate": 1.9516437985067306e-07,
+      "loss": 0.02209635227918625,
+      "mean_token_accuracy": 0.991724469140172,
+      "num_tokens": 46605784.0,
+      "step": 8760
+    },
+    {
+      "entropy": 0.018055033115933837,
+      "epoch": 0.9412395309882747,
+      "grad_norm": 1.4375,
+      "learning_rate": 1.8174982342076819e-07,
+      "loss": 0.02114318758249283,
+      "mean_token_accuracy": 0.9917370740324258,
+      "num_tokens": 46710311.0,
+      "step": 8780
+    },
+    {
+      "entropy": 0.017862834344521163,
+      "epoch": 0.9433835845896147,
+      "grad_norm": 1.40625,
+      "learning_rate": 1.6880862628590343e-07,
+      "loss": 0.017796410620212554,
+      "mean_token_accuracy": 0.9937189962714911,
+      "num_tokens": 46818472.0,
+      "step": 8800
+    },
+    {
+      "entropy": 0.01954027940284959,
+      "epoch": 0.9455276381909548,
+      "grad_norm": 1.6015625,
+      "learning_rate": 1.5634141236947174e-07,
+      "loss": 0.01990315467119217,
+      "mean_token_accuracy": 0.9926852386444807,
+      "num_tokens": 46912066.0,
+      "step": 8820
+    },
+    {
+      "entropy": 0.01788209332662518,
+      "epoch": 0.9476716917922948,
+      "grad_norm": 3.28125,
+      "learning_rate": 1.4434878274309782e-07,
+      "loss": 0.020390361547470093,
+      "mean_token_accuracy": 0.9922173459082841,
+      "num_tokens": 47012839.0,
+      "step": 8840
+    },
+    {
+      "entropy": 0.019225874226185623,
+      "epoch": 0.9498157453936349,
+      "grad_norm": 1.546875,
+      "learning_rate": 1.3283131559766016e-07,
+      "loss": 0.019317436218261718,
+      "mean_token_accuracy": 0.9922500483691692,
+      "num_tokens": 47118537.0,
+      "step": 8860
+    },
+    {
+      "entropy": 0.019644861581696205,
+      "epoch": 0.9519597989949748,
+      "grad_norm": 1.921875,
+      "learning_rate": 1.217895662154178e-07,
+      "loss": 0.02014264166355133,
+      "mean_token_accuracy": 0.9908678252249956,
+      "num_tokens": 47218518.0,
+      "step": 8880
+    },
+    {
+      "entropy": 0.019726947057893084,
+      "epoch": 0.9541038525963149,
+      "grad_norm": 1.859375,
+      "learning_rate": 1.1122406694323628e-07,
+      "loss": 0.021320286393165588,
+      "mean_token_accuracy": 0.9916484046727418,
+      "num_tokens": 47318647.0,
+      "step": 8900
+    },
+    {
+      "entropy": 0.019115359010970678,
+      "epoch": 0.956247906197655,
+      "grad_norm": 1.6875,
+      "learning_rate": 1.0113532716692131e-07,
+      "loss": 0.019532181322574615,
+      "mean_token_accuracy": 0.9924892302602529,
+      "num_tokens": 47423152.0,
+      "step": 8920
+    },
+    {
+      "entropy": 0.019615740292147167,
+      "epoch": 0.9583919597989949,
+      "grad_norm": 1.4921875,
+      "learning_rate": 9.152383328666415e-08,
+      "loss": 0.024791160225868226,
+      "mean_token_accuracy": 0.9896657336503267,
+      "num_tokens": 47525402.0,
+      "step": 8940
+    },
+    {
+      "entropy": 0.018387637558203097,
+      "epoch": 0.960536013400335,
+      "grad_norm": 1.5390625,
+      "learning_rate": 8.239004869358468e-08,
+      "loss": 0.018946020305156706,
+      "mean_token_accuracy": 0.9929318383336068,
+      "num_tokens": 47641223.0,
+      "step": 8960
+    },
+    {
+      "entropy": 0.018579465060270195,
+      "epoch": 0.9626800670016751,
+      "grad_norm": 2.625,
+      "learning_rate": 7.373441374739942e-08,
+      "loss": 0.021176393330097198,
+      "mean_token_accuracy": 0.9917469136416912,
+      "num_tokens": 47742034.0,
+      "step": 8980
+    },
+    {
+      "entropy": 0.019280895775682438,
+      "epoch": 0.964824120603015,
+      "grad_norm": 2.375,
+      "learning_rate": 6.555734575518058e-08,
+      "loss": 0.020877310633659364,
+      "mean_token_accuracy": 0.9919904433190823,
+      "num_tokens": 47836864.0,
+      "step": 9000
+    },
+    {
+      "epoch": 0.964824120603015,
+      "eval_entropy": 0.01948107976220975,
+      "eval_loss": 0.019369935616850853,
+      "eval_mean_token_accuracy": 0.9921530286471049,
+      "eval_num_tokens": 47836864.0,
+      "eval_runtime": 20.3121,
+      "eval_samples_per_second": 36.924,
+      "eval_steps_per_second": 18.462,
+      "step": 9000
+    },
+    {
+      "entropy": 0.01928818029462036,
+      "epoch": 0.9669681742043551,
+      "grad_norm": 2.140625,
+      "learning_rate": 5.7859238951245567e-08,
+      "loss": 0.021238549053668974,
+      "mean_token_accuracy": 0.992226080968976,
+      "num_tokens": 47934338.0,
+      "step": 9020
+    },
+    {
+      "entropy": 0.018642540688506414,
+      "epoch": 0.9691122278056952,
+      "grad_norm": 2.171875,
+      "learning_rate": 5.0640464478142195e-08,
+      "loss": 0.01939709484577179,
+      "mean_token_accuracy": 0.9922053903341294,
+      "num_tokens": 48034410.0,
+      "step": 9040
+    },
+    {
+      "entropy": 0.0197954305883286,
+      "epoch": 0.9712562814070351,
+      "grad_norm": 1.3125,
+      "learning_rate": 4.3901370368764075e-08,
+      "loss": 0.022776152193546294,
+      "mean_token_accuracy": 0.9920450933277607,
+      "num_tokens": 48137325.0,
+      "step": 9060
+    },
+    {
+      "entropy": 0.019472893824354288,
+      "epoch": 0.9734003350083752,
+      "grad_norm": 2.375,
+      "learning_rate": 3.764228152956184e-08,
+      "loss": 0.020811089873313905,
+      "mean_token_accuracy": 0.992926124855876,
+      "num_tokens": 48237086.0,
+      "step": 9080
+    },
+    {
+      "entropy": 0.020447438965675246,
+      "epoch": 0.9755443886097153,
+      "grad_norm": 1.640625,
+      "learning_rate": 3.1863499724887894e-08,
+      "loss": 0.01958942860364914,
+      "mean_token_accuracy": 0.9922197166830301,
+      "num_tokens": 48350555.0,
+      "step": 9100
+    },
+    {
+      "entropy": 0.019909366283036432,
+      "epoch": 0.9776884422110553,
+      "grad_norm": 1.625,
+      "learning_rate": 2.6565303562438027e-08,
+      "loss": 0.020442698895931245,
+      "mean_token_accuracy": 0.9921784479171037,
+      "num_tokens": 48454450.0,
+      "step": 9120
+    },
+    {
+      "entropy": 0.01829673828160594,
+      "epoch": 0.9798324958123953,
+      "grad_norm": 1.59375,
+      "learning_rate": 2.174794847982775e-08,
+      "loss": 0.018291810154914857,
+      "mean_token_accuracy": 0.993328544870019,
+      "num_tokens": 48546425.0,
+      "step": 9140
+    },
+    {
+      "entropy": 0.018716577784925904,
+      "epoch": 0.9819765494137354,
+      "grad_norm": 3.953125,
+      "learning_rate": 1.7411666732271014e-08,
+      "loss": 0.021925486624240875,
+      "mean_token_accuracy": 0.990877429023385,
+      "num_tokens": 48657622.0,
+      "step": 9160
+    },
+    {
+      "entropy": 0.01957869595878492,
+      "epoch": 0.9841206030150754,
+      "grad_norm": 2.03125,
+      "learning_rate": 1.3556667381386946e-08,
+      "loss": 0.020254431664943694,
+      "mean_token_accuracy": 0.992678664624691,
+      "num_tokens": 48758394.0,
+      "step": 9180
+    },
+    {
+      "entropy": 0.020351028728259734,
+      "epoch": 0.9862646566164154,
+      "grad_norm": 2.09375,
+      "learning_rate": 1.0183136285117912e-08,
+      "loss": 0.022582833468914033,
+      "mean_token_accuracy": 0.9906843654811383,
+      "num_tokens": 48864000.0,
+      "step": 9200
+    },
+    {
+      "entropy": 0.019539019266903778,
+      "epoch": 0.9884087102177554,
+      "grad_norm": 2.25,
+      "learning_rate": 7.291236088768916e-09,
+      "loss": 0.01814696490764618,
+      "mean_token_accuracy": 0.9936022497713566,
+      "num_tokens": 48971817.0,
+      "step": 9220
+    },
+    {
+      "entropy": 0.019871230680359987,
+      "epoch": 0.9905527638190955,
+      "grad_norm": 0.9921875,
+      "learning_rate": 4.881106217168308e-09,
+      "loss": 0.021477963030338287,
+      "mean_token_accuracy": 0.991921391710639,
+      "num_tokens": 49081443.0,
+      "step": 9240
+    },
+    {
+      "entropy": 0.02000341287480296,
+      "epoch": 0.9926968174204355,
+      "grad_norm": 2.15625,
+      "learning_rate": 2.9528628679431627e-09,
+      "loss": 0.02385435551404953,
+      "mean_token_accuracy": 0.9907275885343552,
+      "num_tokens": 49184455.0,
+      "step": 9260
+    },
+    {
+      "entropy": 0.019374547952179456,
+      "epoch": 0.9948408710217755,
+      "grad_norm": 2.265625,
+      "learning_rate": 1.5065990059193182e-09,
+      "loss": 0.02085937112569809,
+      "mean_token_accuracy": 0.9919035747647286,
+      "num_tokens": 49283643.0,
+      "step": 9280
+    },
+    {
+      "entropy": 0.01922852643031092,
+      "epoch": 0.9969849246231156,
+      "grad_norm": 1.203125,
+      "learning_rate": 5.423843586371824e-10,
+      "loss": 0.020880751311779022,
+      "mean_token_accuracy": 0.9920044090598822,
+      "num_tokens": 49384046.0,
+      "step": 9300
+    },
+    {
+      "entropy": 0.021174859253278555,
+      "epoch": 0.9991289782244556,
+      "grad_norm": 1.359375,
+      "learning_rate": 6.026541299219979e-11,
+      "loss": 0.020621398091316225,
+      "mean_token_accuracy": 0.9919351786375046,
+      "num_tokens": 49491112.0,
+      "step": 9320
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.019503846389590764,
+      "eval_loss": 0.019395070150494576,
+      "eval_mean_token_accuracy": 0.9922470191319783,
+      "eval_num_tokens": 49533337.0,
+      "eval_runtime": 20.3818,
+      "eval_samples_per_second": 36.798,
+      "eval_steps_per_second": 18.399,
+      "step": 9329
+    }
+  ],
+  "logging_steps": 20,
+  "max_steps": 9329,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 1000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 6.070595342482268e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-9329/training_args.bin b/checkpoint-9329/training_args.bin
new file mode 100644
index 0000000..96cea35
--- /dev/null
+++ b/checkpoint-9329/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:896375d502906677c7c6ac4e94fcf73d56f682781d3d2addb3cca71640f162b7
+size 5777
diff --git a/config.json b/config.json
new file mode 100644
index 0000000..c1ff45d
--- /dev/null
+++ b/config.json
@@ -0,0 +1,63 @@
+{
+  "architectures": [
+    "Qwen3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 40960,
+  "max_window_layers": 28,
+  "model_type": "qwen3",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pad_token_id": null,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 1000000,
+    "rope_type": "default"
+  },
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.7.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
diff --git a/generation_config.json b/generation_config.json
new file mode 100644
index 0000000..d221636
--- /dev/null
+++ b/generation_config.json
@@ -0,0 +1,13 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "temperature": 0.6,
+  "top_k": 20,
+  "top_p": 0.95,
+  "transformers_version": "5.7.0"
+}
diff --git a/model-00001-of-00002.safetensors b/model-00001-of-00002.safetensors
new file mode 100644
index 0000000..9633cab
--- /dev/null
+++ b/model-00001-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:047823c2605576a4620468c2647befc4f2d375aec5619b0b2944fa9b6fffc974
+size 1981418432
diff --git a/model-00002-of-00002.safetensors b/model-00002-of-00002.safetensors
new file mode 100644
index 0000000..72683e9
--- /dev/null
+++ b/model-00002-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e37f5d5ec1fa8eb046faeadd024fb7b509bc791420ee4fada4770fa21b740c9
+size 1459766976
diff --git a/model.safetensors b/model.safetensors
new file mode 100644
index 0000000..feb74a3
--- /dev/null
+++ b/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6e1a2e85a7d6175587f0e8ad7cf78487c1e1689e44b850edd452e28f1b82b6a
+size 3441185608
diff --git a/model.safetensors.index.json b/model.safetensors.index.json
new file mode 100644
index 0000000..f0166d1
--- /dev/null
+++ b/model.safetensors.index.json
@@ -0,0 +1,318 @@
+{
+  "metadata": {
+    "total_parameters": 1720574976,
+    "total_size": 3441149952
+  },
+  "weight_map": {
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.13.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.13.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}
diff --git a/tokenizer.json b/tokenizer.json
new file mode 100644
index 0000000..c7afbed
--- /dev/null
+++ b/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
+size 11422650
diff --git a/tokenizer_config.json b/tokenizer_config.json
new file mode 100644
index 0000000..770e41d
--- /dev/null
+++ b/tokenizer_config.json
@@ -0,0 +1,30 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "local_files_only": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}
diff --git a/training_args.bin b/training_args.bin
new file mode 100644
index 0000000..96cea35
--- /dev/null
+++ b/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:896375d502906677c7c6ac4e94fcf73d56f682781d3d2addb3cca71640f162b7
+size 5777