Upload checkpoint checkpoint-200

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* 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-200/tokenizer.json filter=lfs diff=lfs merge=lfs -text

checkpoint-200/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoint-200/config.json

@@ -0,0 +1,30 @@
+{
+  "_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 896,
+  "initializer_range": 0.02,
+  "intermediate_size": 4864,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 24,
+  "model_type": "qwen2",
+  "num_attention_heads": 14,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 2,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": false,
+  "use_mrope": false,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}

checkpoint-200/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
+  "transformers_version": "4.49.0"
+}

checkpoint-200/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e1bbe94700bcac026a8373a6bd550cd57ff23c2ad7effe2cc1db899ffc7260a
+size 988097824

checkpoint-200/special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_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|>"
+  ],
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-200/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5eee858c5123a4279c3e1f7b81247343f356ac767940b2692a928ad929543214
+size 11422063

checkpoint-200/tokenizer_config.json

@@ -0,0 +1,209 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_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|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# 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>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\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\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoint-200/trainer_state.json

@@ -0,0 +1,2833 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.3931203931203931,
+  "eval_steps": 500,
+  "global_step": 200,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 562.125,
+      "epoch": 0.0019656019656019656,
+      "grad_norm": 0.0,
+      "kl": 0.0,
+      "learning_rate": 5e-08,
+      "loss": 0.0,
+      "num_tokens": 22458.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 1
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 382.84375,
+      "epoch": 0.003931203931203931,
+      "grad_norm": 0.0,
+      "kl": 0.0,
+      "learning_rate": 1e-07,
+      "loss": 0.0,
+      "num_tokens": 39539.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 2
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 296.84375,
+      "epoch": 0.005896805896805897,
+      "grad_norm": 0.0,
+      "kl": 0.0,
+      "learning_rate": 1.5e-07,
+      "loss": 0.0,
+      "num_tokens": 52892.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 3
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 562.84375,
+      "epoch": 0.007862407862407862,
+      "grad_norm": 0.0,
+      "kl": 0.0,
+      "learning_rate": 2e-07,
+      "loss": 0.0,
+      "num_tokens": 73799.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 4
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 454.9375,
+      "epoch": 0.009828009828009828,
+      "grad_norm": 0.0,
+      "kl": 0.0,
+      "learning_rate": 2.5e-07,
+      "loss": 0.0,
+      "num_tokens": 92389.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 5
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 401.5625,
+      "epoch": 0.011793611793611793,
+      "grad_norm": 0.8271026359027706,
+      "kl": 0.0,
+      "learning_rate": 3e-07,
+      "loss": 0.0459,
+      "num_tokens": 108305.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 6
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 390.625,
+      "epoch": 0.013759213759213759,
+      "grad_norm": 0.0003446225536382493,
+      "kl": 0.0002499818801879883,
+      "learning_rate": 3.5e-07,
+      "loss": 0.0,
+      "num_tokens": 125607.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 7
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 448.75,
+      "epoch": 0.015724815724815724,
+      "grad_norm": 0.0013026212429223525,
+      "kl": 0.0004665851593017578,
+      "learning_rate": 4e-07,
+      "loss": 0.0,
+      "num_tokens": 142955.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 8
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 461.90625,
+      "epoch": 0.01769041769041769,
+      "grad_norm": 0.0021171088963702245,
+      "kl": 0.0006817877292633057,
+      "learning_rate": 4.5e-07,
+      "loss": 0.0,
+      "num_tokens": 162956.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 9
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 330.6875,
+      "epoch": 0.019656019656019656,
+      "grad_norm": 1.3023849518044601,
+      "kl": 0.0006229877471923828,
+      "learning_rate": 5e-07,
+      "loss": 0.0444,
+      "num_tokens": 177758.0,
+      "reward": 0.0625,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.0625,
+      "step": 10
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 513.1875,
+      "epoch": 0.021621621621621623,
+      "grad_norm": 0.00027502974416413834,
+      "kl": 0.00018683075904846191,
+      "learning_rate": 5.5e-07,
+      "loss": 0.0,
+      "num_tokens": 197364.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 11
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 424.625,
+      "epoch": 0.023587223587223587,
+      "grad_norm": 0.6170315097653724,
+      "kl": 0.00029702484607696533,
+      "learning_rate": 6e-07,
+      "loss": -0.0489,
+      "num_tokens": 214156.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 12
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 532.1875,
+      "epoch": 0.025552825552825554,
+      "grad_norm": 0.0002590285678817397,
+      "kl": 0.00019553303718566895,
+      "learning_rate": 6.5e-07,
+      "loss": 0.0,
+      "num_tokens": 234938.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 13
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 391.6875,
+      "epoch": 0.027518427518427518,
+      "grad_norm": 0.2468481109530296,
+      "kl": 0.0002957582473754883,
+      "learning_rate": 7e-07,
+      "loss": -0.042,
+      "num_tokens": 250970.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 14
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 526.75,
+      "epoch": 0.029484029484029485,
+      "grad_norm": 0.6658783590054981,
+      "kl": 0.00033867359161376953,
+      "learning_rate": 7.5e-07,
+      "loss": -0.0001,
+      "num_tokens": 271842.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 15
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 357.5,
+      "epoch": 0.03144963144963145,
+      "grad_norm": 0.001449294133896928,
+      "kl": 0.00034862756729125977,
+      "learning_rate": 8e-07,
+      "loss": 0.0,
+      "num_tokens": 288920.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 16
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 466.9375,
+      "epoch": 0.03341523341523341,
+      "grad_norm": 0.000550624611359764,
+      "kl": 0.00036454200744628906,
+      "learning_rate": 8.499999999999999e-07,
+      "loss": 0.0,
+      "num_tokens": 306990.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 17
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 425.15625,
+      "epoch": 0.03538083538083538,
+      "grad_norm": 0.4966139299418229,
+      "kl": 0.0004194974899291992,
+      "learning_rate": 9e-07,
+      "loss": -0.059,
+      "num_tokens": 324483.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 18
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 558.84375,
+      "epoch": 0.03734643734643735,
+      "grad_norm": 0.0005133453694281673,
+      "kl": 0.000286102294921875,
+      "learning_rate": 9.499999999999999e-07,
+      "loss": 0.0,
+      "num_tokens": 346726.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 19
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 462.96875,
+      "epoch": 0.03931203931203931,
+      "grad_norm": 0.7524081119290347,
+      "kl": 0.0004723072052001953,
+      "learning_rate": 1e-06,
+      "loss": -0.051,
+      "num_tokens": 364915.0,
+      "reward": 0.0625,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.0625,
+      "step": 20
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 442.84375,
+      "epoch": 0.041277641277641275,
+      "grad_norm": 0.002276209851987083,
+      "kl": 0.00047028064727783203,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 382538.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 21
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 487.8125,
+      "epoch": 0.043243243243243246,
+      "grad_norm": 0.0009159776748314612,
+      "kl": 0.0003075599670410156,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 402660.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 22
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 521.09375,
+      "epoch": 0.04520884520884521,
+      "grad_norm": 0.4431344945845042,
+      "kl": 0.0008206367492675781,
+      "learning_rate": 1e-06,
+      "loss": 0.0654,
+      "num_tokens": 423169.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 23
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 424.9375,
+      "epoch": 0.04717444717444717,
+      "grad_norm": 0.00201223494239014,
+      "kl": 0.0007357597351074219,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 439839.0,
+      "reward": 0.0625,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0625,
+      "step": 24
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 593.625,
+      "epoch": 0.04914004914004914,
+      "grad_norm": 0.18561111784701523,
+      "kl": 0.0019412040710449219,
+      "learning_rate": 1e-06,
+      "loss": -0.0044,
+      "num_tokens": 463265.0,
+      "reward": 0.09375,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.09375,
+      "step": 25
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 280.375,
+      "epoch": 0.05110565110565111,
+      "grad_norm": 5.884242558131956,
+      "kl": 0.0025135278701782227,
+      "learning_rate": 1e-06,
+      "loss": -0.0569,
+      "num_tokens": 475453.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 26
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 513.90625,
+      "epoch": 0.05307125307125307,
+      "grad_norm": 0.0019098622203968013,
+      "kl": 0.0006109476089477539,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 495570.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 27
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 537.8125,
+      "epoch": 0.055036855036855035,
+      "grad_norm": 0.22075214354309136,
+      "kl": 0.0009493827819824219,
+      "learning_rate": 1e-06,
+      "loss": -0.0333,
+      "num_tokens": 516232.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 28
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 485.6875,
+      "epoch": 0.057002457002457,
+      "grad_norm": 0.0027609914881789005,
+      "kl": 0.0013211965560913086,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 534828.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 29
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 394.8125,
+      "epoch": 0.05896805896805897,
+      "grad_norm": 1035336.9697758019,
+      "kl": 7488.001363635063,
+      "learning_rate": 1e-06,
+      "loss": 7.5002,
+      "num_tokens": 550682.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 30
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 472.96875,
+      "epoch": 0.060933660933660934,
+      "grad_norm": 0.005787911590971138,
+      "kl": 0.002418994903564453,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 568901.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 31
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 532.1875,
+      "epoch": 0.0628992628992629,
+      "grad_norm": 0.22606153101175416,
+      "kl": 0.0019965171813964844,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 589439.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 32
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 443.78125,
+      "epoch": 0.06486486486486487,
+      "grad_norm": 0.35741687124698773,
+      "kl": 0.002090930938720703,
+      "learning_rate": 1e-06,
+      "loss": 0.007,
+      "num_tokens": 606976.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 33
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 582.25,
+      "epoch": 0.06683046683046683,
+      "grad_norm": 0.0012562903075668153,
+      "kl": 0.0010848045349121094,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 629146.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 34
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 445.96875,
+      "epoch": 0.0687960687960688,
+      "grad_norm": 0.9035998758815753,
+      "kl": 0.0037384033203125,
+      "learning_rate": 1e-06,
+      "loss": 0.0705,
+      "num_tokens": 646643.0,
+      "reward": 0.125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.125,
+      "step": 35
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 444.6875,
+      "epoch": 0.07076167076167077,
+      "grad_norm": 0.3422524583969719,
+      "kl": 0.0064411163330078125,
+      "learning_rate": 1e-06,
+      "loss": -0.0749,
+      "num_tokens": 666223.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 36
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 375.15625,
+      "epoch": 0.07272727272727272,
+      "grad_norm": 0.6543309959643355,
+      "kl": 0.025150299072265625,
+      "learning_rate": 1e-06,
+      "loss": -0.074,
+      "num_tokens": 681430.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 37
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 465.40625,
+      "epoch": 0.0746928746928747,
+      "grad_norm": 0.4629932120108371,
+      "kl": 0.006771087646484375,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 699985.0,
+      "reward": 0.09375,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.09375,
+      "step": 38
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 386.46875,
+      "epoch": 0.07665847665847667,
+      "grad_norm": 0.2536795642942375,
+      "kl": 0.010662078857421875,
+      "learning_rate": 1e-06,
+      "loss": -0.0128,
+      "num_tokens": 715666.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 39
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 398.90625,
+      "epoch": 0.07862407862407862,
+      "grad_norm": 0.36055266191548657,
+      "kl": 0.014057159423828125,
+      "learning_rate": 1e-06,
+      "loss": -0.0222,
+      "num_tokens": 732739.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 40
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 497.78125,
+      "epoch": 0.08058968058968059,
+      "grad_norm": 0.5439508463238408,
+      "kl": 0.011962890625,
+      "learning_rate": 1e-06,
+      "loss": 0.0736,
+      "num_tokens": 752036.0,
+      "reward": 0.0625,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.0625,
+      "step": 41
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 553.9375,
+      "epoch": 0.08255528255528255,
+      "grad_norm": 0.2940020586639228,
+      "kl": 0.01055145263671875,
+      "learning_rate": 1e-06,
+      "loss": -0.0098,
+      "num_tokens": 774304.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 42
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 506.78125,
+      "epoch": 0.08452088452088452,
+      "grad_norm": 0.34987523935436315,
+      "kl": 0.022228240966796875,
+      "learning_rate": 1e-06,
+      "loss": 0.0136,
+      "num_tokens": 794573.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 43
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 534.71875,
+      "epoch": 0.08648648648648649,
+      "grad_norm": 0.28775109044810143,
+      "kl": 0.016704559326171875,
+      "learning_rate": 1e-06,
+      "loss": -0.022,
+      "num_tokens": 814952.0,
+      "reward": 0.0625,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.0625,
+      "step": 44
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 551.125,
+      "epoch": 0.08845208845208845,
+      "grad_norm": 0.5294418172877363,
+      "kl": 0.01171875,
+      "learning_rate": 1e-06,
+      "loss": 0.0127,
+      "num_tokens": 837090.0,
+      "reward": 0.0625,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.0625,
+      "step": 45
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 424.09375,
+      "epoch": 0.09041769041769042,
+      "grad_norm": 0.7442762340799661,
+      "kl": 0.0166015625,
+      "learning_rate": 1e-06,
+      "loss": -0.0527,
+      "num_tokens": 855063.0,
+      "reward": 0.28125,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.28125,
+      "step": 46
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 578.90625,
+      "epoch": 0.09238329238329239,
+      "grad_norm": 0.9964430347064654,
+      "kl": 0.01439666748046875,
+      "learning_rate": 1e-06,
+      "loss": -0.0347,
+      "num_tokens": 876960.0,
+      "reward": 0.1875,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.1875,
+      "step": 47
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 516.65625,
+      "epoch": 0.09434889434889435,
+      "grad_norm": 0.0029736329333483326,
+      "kl": 0.015178680419921875,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 897477.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 48
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 406.53125,
+      "epoch": 0.09631449631449632,
+      "grad_norm": 0.42366305164369583,
+      "kl": 0.0214385986328125,
+      "learning_rate": 1e-06,
+      "loss": 0.023,
+      "num_tokens": 913848.0,
+      "reward": 0.0625,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.0625,
+      "step": 49
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 615.46875,
+      "epoch": 0.09828009828009827,
+      "grad_norm": 0.2894821606425995,
+      "kl": 0.010334014892578125,
+      "learning_rate": 1e-06,
+      "loss": -0.0008,
+      "num_tokens": 936513.0,
+      "reward": 0.125,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.125,
+      "step": 50
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 484.40625,
+      "epoch": 0.10024570024570024,
+      "grad_norm": 0.8466055231288873,
+      "kl": 0.01735687255859375,
+      "learning_rate": 1e-06,
+      "loss": 0.0053,
+      "num_tokens": 955662.0,
+      "reward": 0.125,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.125,
+      "step": 51
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 603.40625,
+      "epoch": 0.10221130221130222,
+      "grad_norm": 0.4594988801570104,
+      "kl": 0.01914215087890625,
+      "learning_rate": 1e-06,
+      "loss": 0.0157,
+      "num_tokens": 978603.0,
+      "reward": 0.0625,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.0625,
+      "step": 52
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 571.53125,
+      "epoch": 0.10417690417690417,
+      "grad_norm": 0.6301487761037352,
+      "kl": 0.01680755615234375,
+      "learning_rate": 1e-06,
+      "loss": -0.0518,
+      "num_tokens": 1000414.0,
+      "reward": 0.125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.125,
+      "step": 53
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 480.625,
+      "epoch": 0.10614250614250614,
+      "grad_norm": 0.1705061403375656,
+      "kl": 0.019412994384765625,
+      "learning_rate": 1e-06,
+      "loss": -0.0151,
+      "num_tokens": 1019984.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 54
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 586.5625,
+      "epoch": 0.10810810810810811,
+      "grad_norm": 0.3563070736118851,
+      "kl": 0.01416778564453125,
+      "learning_rate": 1e-06,
+      "loss": -0.0113,
+      "num_tokens": 1043134.0,
+      "reward": 0.03125,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.03125,
+      "step": 55
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 573.78125,
+      "epoch": 0.11007371007371007,
+      "grad_norm": 0.8170380900607339,
+      "kl": 0.0185394287109375,
+      "learning_rate": 1e-06,
+      "loss": 0.0107,
+      "num_tokens": 1065473.0,
+      "reward": 0.125,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.125,
+      "step": 56
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 596.25,
+      "epoch": 0.11203931203931204,
+      "grad_norm": 0.6138181155074318,
+      "kl": 0.01763916015625,
+      "learning_rate": 1e-06,
+      "loss": -0.0173,
+      "num_tokens": 1088233.0,
+      "reward": 0.0625,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.0625,
+      "step": 57
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 579.96875,
+      "epoch": 0.114004914004914,
+      "grad_norm": 0.5592170560306685,
+      "kl": 0.015392303466796875,
+      "learning_rate": 1e-06,
+      "loss": -0.042,
+      "num_tokens": 1111490.0,
+      "reward": 0.125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.125,
+      "step": 58
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 590.3125,
+      "epoch": 0.11597051597051597,
+      "grad_norm": 0.004534498568535779,
+      "kl": 0.017578125,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 1134000.0,
+      "reward": 0.125,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.125,
+      "step": 59
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 552.75,
+      "epoch": 0.11793611793611794,
+      "grad_norm": 0.26122008088795556,
+      "kl": 0.019062042236328125,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 1155236.0,
+      "reward": 0.09375,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.09375,
+      "step": 60
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 533.46875,
+      "epoch": 0.1199017199017199,
+      "grad_norm": 0.6246443042334695,
+      "kl": 0.01920318603515625,
+      "learning_rate": 1e-06,
+      "loss": -0.0014,
+      "num_tokens": 1175617.0,
+      "reward": 0.09375,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.09375,
+      "step": 61
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 484.34375,
+      "epoch": 0.12186732186732187,
+      "grad_norm": 1.0553043786365885,
+      "kl": 0.0269622802734375,
+      "learning_rate": 1e-06,
+      "loss": 0.0461,
+      "num_tokens": 1194332.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 62
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 604.84375,
+      "epoch": 0.12383292383292384,
+      "grad_norm": 0.5614007392931322,
+      "kl": 0.0162506103515625,
+      "learning_rate": 1e-06,
+      "loss": 0.0098,
+      "num_tokens": 1217363.0,
+      "reward": 0.1875,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.1875,
+      "step": 63
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 579.59375,
+      "epoch": 0.1257985257985258,
+      "grad_norm": 0.001324981221462767,
+      "kl": 0.0164794921875,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 1239884.0,
+      "reward": 0.0625,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0625,
+      "step": 64
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 643.0,
+      "epoch": 0.12776412776412777,
+      "grad_norm": 0.6061011991202276,
+      "kl": 0.01563262939453125,
+      "learning_rate": 1e-06,
+      "loss": -0.065,
+      "num_tokens": 1263644.0,
+      "reward": 0.125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.125,
+      "step": 65
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 487.875,
+      "epoch": 0.12972972972972974,
+      "grad_norm": 0.8366609324695296,
+      "kl": 0.02077484130859375,
+      "learning_rate": 1e-06,
+      "loss": 0.0478,
+      "num_tokens": 1282820.0,
+      "reward": 0.15625,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.15625,
+      "step": 66
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 662.59375,
+      "epoch": 0.1316953316953317,
+      "grad_norm": 0.17490281761649612,
+      "kl": 0.0147705078125,
+      "learning_rate": 1e-06,
+      "loss": 0.0204,
+      "num_tokens": 1307457.0,
+      "reward": 0.09375,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.09375,
+      "step": 67
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 525.03125,
+      "epoch": 0.13366093366093365,
+      "grad_norm": 0.8215530702995869,
+      "kl": 0.020721435546875,
+      "learning_rate": 1e-06,
+      "loss": 0.064,
+      "num_tokens": 1327798.0,
+      "reward": 0.25,
+      "reward_std": 0.3535533845424652,
+      "rewards/reward_func": 0.25,
+      "step": 68
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 567.1875,
+      "epoch": 0.13562653562653562,
+      "grad_norm": 0.65089823105939,
+      "kl": 0.019927978515625,
+      "learning_rate": 1e-06,
+      "loss": -0.0026,
+      "num_tokens": 1350334.0,
+      "reward": 0.15625,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.15625,
+      "step": 69
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 584.78125,
+      "epoch": 0.1375921375921376,
+      "grad_norm": 1.008422919649498,
+      "kl": 0.01912689208984375,
+      "learning_rate": 1e-06,
+      "loss": 0.0025,
+      "num_tokens": 1372311.0,
+      "reward": 0.25,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.25,
+      "step": 70
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 545.125,
+      "epoch": 0.13955773955773956,
+      "grad_norm": 0.6752475010590075,
+      "kl": 0.01819610595703125,
+      "learning_rate": 1e-06,
+      "loss": 0.0177,
+      "num_tokens": 1393645.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 71
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 526.40625,
+      "epoch": 0.14152334152334153,
+      "grad_norm": 0.7868187210048034,
+      "kl": 0.0257568359375,
+      "learning_rate": 1e-06,
+      "loss": -0.0076,
+      "num_tokens": 1414260.0,
+      "reward": 0.15625,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.15625,
+      "step": 72
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 465.46875,
+      "epoch": 0.14348894348894348,
+      "grad_norm": 0.40424658361331584,
+      "kl": 0.02496337890625,
+      "learning_rate": 1e-06,
+      "loss": 0.0034,
+      "num_tokens": 1432617.0,
+      "reward": 0.09375,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.09375,
+      "step": 73
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 523.0,
+      "epoch": 0.14545454545454545,
+      "grad_norm": 0.7217516854340651,
+      "kl": 0.0191497802734375,
+      "learning_rate": 1e-06,
+      "loss": -0.0438,
+      "num_tokens": 1452631.0,
+      "reward": 0.125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.125,
+      "step": 74
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 564.59375,
+      "epoch": 0.14742014742014742,
+      "grad_norm": 1.020886020177138,
+      "kl": 0.02172088623046875,
+      "learning_rate": 1e-06,
+      "loss": 0.0443,
+      "num_tokens": 1474180.0,
+      "reward": 0.25,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.25,
+      "step": 75
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 599.03125,
+      "epoch": 0.1493857493857494,
+      "grad_norm": 0.38779629766219476,
+      "kl": 0.021942138671875,
+      "learning_rate": 1e-06,
+      "loss": 0.0253,
+      "num_tokens": 1496597.0,
+      "reward": 0.0625,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.0625,
+      "step": 76
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 535.65625,
+      "epoch": 0.15135135135135136,
+      "grad_norm": 0.5231239100329028,
+      "kl": 0.0213470458984375,
+      "learning_rate": 1e-06,
+      "loss": -0.0072,
+      "num_tokens": 1518002.0,
+      "reward": 0.09375,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.09375,
+      "step": 77
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 483.28125,
+      "epoch": 0.15331695331695333,
+      "grad_norm": 1.01964836669152,
+      "kl": 0.02301025390625,
+      "learning_rate": 1e-06,
+      "loss": 0.0253,
+      "num_tokens": 1538251.0,
+      "reward": 0.1875,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.1875,
+      "step": 78
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 597.21875,
+      "epoch": 0.15528255528255527,
+      "grad_norm": 0.3894010167090417,
+      "kl": 0.01868438720703125,
+      "learning_rate": 1e-06,
+      "loss": 0.0049,
+      "num_tokens": 1561066.0,
+      "reward": 0.125,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.125,
+      "step": 79
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 494.21875,
+      "epoch": 0.15724815724815724,
+      "grad_norm": 0.4559857762776204,
+      "kl": 0.02178955078125,
+      "learning_rate": 1e-06,
+      "loss": -0.0144,
+      "num_tokens": 1579851.0,
+      "reward": 0.125,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.125,
+      "step": 80
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 459.4375,
+      "epoch": 0.15921375921375922,
+      "grad_norm": 0.7408328801166559,
+      "kl": 0.02642822265625,
+      "learning_rate": 1e-06,
+      "loss": 0.0034,
+      "num_tokens": 1597675.0,
+      "reward": 0.15625,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.15625,
+      "step": 81
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 601.90625,
+      "epoch": 0.16117936117936119,
+      "grad_norm": 0.4202464041477536,
+      "kl": 0.01740264892578125,
+      "learning_rate": 1e-06,
+      "loss": 0.0084,
+      "num_tokens": 1621632.0,
+      "reward": 0.125,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.125,
+      "step": 82
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 517.21875,
+      "epoch": 0.16314496314496316,
+      "grad_norm": 0.7785902853786559,
+      "kl": 0.02386474609375,
+      "learning_rate": 1e-06,
+      "loss": 0.0036,
+      "num_tokens": 1641873.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 83
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 390.6875,
+      "epoch": 0.1651105651105651,
+      "grad_norm": 0.6791841008394204,
+      "kl": 0.034637451171875,
+      "learning_rate": 1e-06,
+      "loss": -0.0184,
+      "num_tokens": 1658453.0,
+      "reward": 0.1875,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.1875,
+      "step": 84
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 485.78125,
+      "epoch": 0.16707616707616707,
+      "grad_norm": 0.4367719984766766,
+      "kl": 0.0251617431640625,
+      "learning_rate": 1e-06,
+      "loss": 0.0156,
+      "num_tokens": 1676992.0,
+      "reward": 0.125,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.125,
+      "step": 85
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 528.875,
+      "epoch": 0.16904176904176904,
+      "grad_norm": 0.7222029349504244,
+      "kl": 0.02223968505859375,
+      "learning_rate": 1e-06,
+      "loss": 0.0459,
+      "num_tokens": 1700584.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 86
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 658.125,
+      "epoch": 0.171007371007371,
+      "grad_norm": 0.34834776108087606,
+      "kl": 0.01824951171875,
+      "learning_rate": 1e-06,
+      "loss": 0.0326,
+      "num_tokens": 1725524.0,
+      "reward": 0.125,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.125,
+      "step": 87
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 644.5625,
+      "epoch": 0.17297297297297298,
+      "grad_norm": 0.6750518830704131,
+      "kl": 0.01828765869140625,
+      "learning_rate": 1e-06,
+      "loss": 0.0412,
+      "num_tokens": 1749556.0,
+      "reward": 0.15625,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.15625,
+      "step": 88
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 547.15625,
+      "epoch": 0.17493857493857493,
+      "grad_norm": 0.31981853904702845,
+      "kl": 0.01940155029296875,
+      "learning_rate": 1e-06,
+      "loss": 0.012,
+      "num_tokens": 1770175.0,
+      "reward": 0.1875,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.1875,
+      "step": 89
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 475.75,
+      "epoch": 0.1769041769041769,
+      "grad_norm": 1.176119812297408,
+      "kl": 0.0265045166015625,
+      "learning_rate": 1e-06,
+      "loss": -0.0174,
+      "num_tokens": 1789467.0,
+      "reward": 0.1875,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.1875,
+      "step": 90
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 448.625,
+      "epoch": 0.17886977886977887,
+      "grad_norm": 0.7247013137515755,
+      "kl": 0.02973175048828125,
+      "learning_rate": 1e-06,
+      "loss": 0.0025,
+      "num_tokens": 1807295.0,
+      "reward": 0.09375,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.09375,
+      "step": 91
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 461.84375,
+      "epoch": 0.18083538083538084,
+      "grad_norm": 0.9576563539486035,
+      "kl": 0.027984619140625,
+      "learning_rate": 1e-06,
+      "loss": 0.0572,
+      "num_tokens": 1827104.0,
+      "reward": 0.28125,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.28125,
+      "step": 92
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 549.375,
+      "epoch": 0.1828009828009828,
+      "grad_norm": 0.8080609769164848,
+      "kl": 0.02375030517578125,
+      "learning_rate": 1e-06,
+      "loss": 0.0641,
+      "num_tokens": 1848272.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 93
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 673.53125,
+      "epoch": 0.18476658476658478,
+      "grad_norm": 0.6070813087390035,
+      "kl": 0.0193634033203125,
+      "learning_rate": 1e-06,
+      "loss": 0.0232,
+      "num_tokens": 1873533.0,
+      "reward": 0.125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.125,
+      "step": 94
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 582.75,
+      "epoch": 0.18673218673218672,
+      "grad_norm": 0.6479257231531951,
+      "kl": 0.0185394287109375,
+      "learning_rate": 1e-06,
+      "loss": 0.0017,
+      "num_tokens": 1895601.0,
+      "reward": 0.3125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.3125,
+      "step": 95
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 443.5,
+      "epoch": 0.1886977886977887,
+      "grad_norm": 0.7375270346991556,
+      "kl": 0.0291290283203125,
+      "learning_rate": 1e-06,
+      "loss": 0.0343,
+      "num_tokens": 1912993.0,
+      "reward": 0.15625,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.15625,
+      "step": 96
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 499.125,
+      "epoch": 0.19066339066339066,
+      "grad_norm": 0.6426005454499798,
+      "kl": 0.0281524658203125,
+      "learning_rate": 1e-06,
+      "loss": -0.0165,
+      "num_tokens": 1933631.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 97
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 583.21875,
+      "epoch": 0.19262899262899263,
+      "grad_norm": 0.656305823795339,
+      "kl": 0.0197601318359375,
+      "learning_rate": 1e-06,
+      "loss": 0.0098,
+      "num_tokens": 1955588.0,
+      "reward": 0.125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.125,
+      "step": 98
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 604.5,
+      "epoch": 0.1945945945945946,
+      "grad_norm": 0.7159925904721693,
+      "kl": 0.0202484130859375,
+      "learning_rate": 1e-06,
+      "loss": -0.0024,
+      "num_tokens": 1978466.0,
+      "reward": 0.15625,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.15625,
+      "step": 99
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 623.8125,
+      "epoch": 0.19656019656019655,
+      "grad_norm": 0.6011532165345448,
+      "kl": 0.0197296142578125,
+      "learning_rate": 1e-06,
+      "loss": 0.015,
+      "num_tokens": 2001554.0,
+      "reward": 0.1875,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.1875,
+      "step": 100
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 487.84375,
+      "epoch": 0.19852579852579852,
+      "grad_norm": 0.6696267166180018,
+      "kl": 0.02446746826171875,
+      "learning_rate": 1e-06,
+      "loss": 0.0323,
+      "num_tokens": 2020681.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 101
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 512.90625,
+      "epoch": 0.2004914004914005,
+      "grad_norm": 0.5144645053944626,
+      "kl": 0.0222320556640625,
+      "learning_rate": 1e-06,
+      "loss": 0.0308,
+      "num_tokens": 2040058.0,
+      "reward": 0.0625,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.0625,
+      "step": 102
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 492.59375,
+      "epoch": 0.20245700245700246,
+      "grad_norm": 0.9792203746023931,
+      "kl": 0.02777099609375,
+      "learning_rate": 1e-06,
+      "loss": 0.0185,
+      "num_tokens": 2059481.0,
+      "reward": 0.28125,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.28125,
+      "step": 103
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 540.5625,
+      "epoch": 0.20442260442260443,
+      "grad_norm": 0.6773524274596018,
+      "kl": 0.0245513916015625,
+      "learning_rate": 1e-06,
+      "loss": 0.0115,
+      "num_tokens": 2079795.0,
+      "reward": 0.125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.125,
+      "step": 104
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 629.75,
+      "epoch": 0.20638820638820637,
+      "grad_norm": 0.0014408090990997875,
+      "kl": 0.020843505859375,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 2104281.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/reward_func": 0.0,
+      "step": 105
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 489.625,
+      "epoch": 0.20835380835380835,
+      "grad_norm": 0.43577632912535613,
+      "kl": 0.02980804443359375,
+      "learning_rate": 1e-06,
+      "loss": 0.0265,
+      "num_tokens": 2124275.0,
+      "reward": 0.25,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.25,
+      "step": 106
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 480.90625,
+      "epoch": 0.21031941031941032,
+      "grad_norm": 0.5136803725458018,
+      "kl": 0.02611541748046875,
+      "learning_rate": 1e-06,
+      "loss": 0.0184,
+      "num_tokens": 2143080.0,
+      "reward": 0.3125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.3125,
+      "step": 107
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 659.9375,
+      "epoch": 0.2122850122850123,
+      "grad_norm": 0.4897037723304078,
+      "kl": 0.0196380615234375,
+      "learning_rate": 1e-06,
+      "loss": -0.0062,
+      "num_tokens": 2167550.0,
+      "reward": 0.15625,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.15625,
+      "step": 108
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 549.53125,
+      "epoch": 0.21425061425061426,
+      "grad_norm": 0.6823729897684875,
+      "kl": 0.020660400390625,
+      "learning_rate": 1e-06,
+      "loss": 0.0027,
+      "num_tokens": 2188815.0,
+      "reward": 0.1875,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.1875,
+      "step": 109
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 492.4375,
+      "epoch": 0.21621621621621623,
+      "grad_norm": 0.8520053469926316,
+      "kl": 0.02719879150390625,
+      "learning_rate": 1e-06,
+      "loss": 0.0077,
+      "num_tokens": 2207977.0,
+      "reward": 0.28125,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.28125,
+      "step": 110
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 487.34375,
+      "epoch": 0.21818181818181817,
+      "grad_norm": 0.8983378201666078,
+      "kl": 0.024871826171875,
+      "learning_rate": 1e-06,
+      "loss": 0.0207,
+      "num_tokens": 2226818.0,
+      "reward": 0.40625,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.40625,
+      "step": 111
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 534.71875,
+      "epoch": 0.22014742014742014,
+      "grad_norm": 0.6341326642846564,
+      "kl": 0.022857666015625,
+      "learning_rate": 1e-06,
+      "loss": 0.005,
+      "num_tokens": 2247351.0,
+      "reward": 0.25,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.25,
+      "step": 112
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 509.96875,
+      "epoch": 0.2221130221130221,
+      "grad_norm": 0.5084014496191064,
+      "kl": 0.02394866943359375,
+      "learning_rate": 1e-06,
+      "loss": 0.0121,
+      "num_tokens": 2267646.0,
+      "reward": 0.1875,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.1875,
+      "step": 113
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 493.46875,
+      "epoch": 0.22407862407862408,
+      "grad_norm": 0.47081172119878306,
+      "kl": 0.023681640625,
+      "learning_rate": 1e-06,
+      "loss": 0.003,
+      "num_tokens": 2287461.0,
+      "reward": 0.21875,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.21875,
+      "step": 114
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 593.46875,
+      "epoch": 0.22604422604422605,
+      "grad_norm": 0.5402745390200868,
+      "kl": 0.0226593017578125,
+      "learning_rate": 1e-06,
+      "loss": 0.0254,
+      "num_tokens": 2310300.0,
+      "reward": 0.21875,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.21875,
+      "step": 115
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 659.34375,
+      "epoch": 0.228009828009828,
+      "grad_norm": 0.7384552819583179,
+      "kl": 0.01863861083984375,
+      "learning_rate": 1e-06,
+      "loss": -0.011,
+      "num_tokens": 2336289.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 116
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 426.0625,
+      "epoch": 0.22997542997542997,
+      "grad_norm": 0.9856793371703316,
+      "kl": 0.029998779296875,
+      "learning_rate": 1e-06,
+      "loss": 0.0107,
+      "num_tokens": 2353069.0,
+      "reward": 0.1875,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.1875,
+      "step": 117
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 553.15625,
+      "epoch": 0.23194103194103194,
+      "grad_norm": 0.6524109632549928,
+      "kl": 0.0211334228515625,
+      "learning_rate": 1e-06,
+      "loss": 0.0166,
+      "num_tokens": 2374500.0,
+      "reward": 0.25,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.25,
+      "step": 118
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 495.96875,
+      "epoch": 0.2339066339066339,
+      "grad_norm": 0.750164710998597,
+      "kl": 0.0268707275390625,
+      "learning_rate": 1e-06,
+      "loss": -0.0291,
+      "num_tokens": 2393455.0,
+      "reward": 0.34375,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.34375,
+      "step": 119
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 519.375,
+      "epoch": 0.23587223587223588,
+      "grad_norm": 0.8136338780518924,
+      "kl": 0.0233917236328125,
+      "learning_rate": 1e-06,
+      "loss": 0.0213,
+      "num_tokens": 2413821.0,
+      "reward": 0.25,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.25,
+      "step": 120
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 492.5,
+      "epoch": 0.23783783783783785,
+      "grad_norm": 0.9472454617054751,
+      "kl": 0.0245513916015625,
+      "learning_rate": 1e-06,
+      "loss": 0.0434,
+      "num_tokens": 2432727.0,
+      "reward": 0.3125,
+      "reward_std": 0.3535533845424652,
+      "rewards/reward_func": 0.3125,
+      "step": 121
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 650.125,
+      "epoch": 0.2398034398034398,
+      "grad_norm": 0.7941644508393335,
+      "kl": 0.0188446044921875,
+      "learning_rate": 1e-06,
+      "loss": 0.0161,
+      "num_tokens": 2457059.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 122
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 505.40625,
+      "epoch": 0.24176904176904176,
+      "grad_norm": 0.9756007077675318,
+      "kl": 0.02396392822265625,
+      "learning_rate": 1e-06,
+      "loss": 0.0303,
+      "num_tokens": 2477004.0,
+      "reward": 0.53125,
+      "reward_std": 0.30935921147465706,
+      "rewards/reward_func": 0.53125,
+      "step": 123
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 483.96875,
+      "epoch": 0.24373464373464374,
+      "grad_norm": 0.6941208484996456,
+      "kl": 0.027313232421875,
+      "learning_rate": 1e-06,
+      "loss": 0.0341,
+      "num_tokens": 2495959.0,
+      "reward": 0.40625,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.40625,
+      "step": 124
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 561.3125,
+      "epoch": 0.2457002457002457,
+      "grad_norm": 0.5265733548592475,
+      "kl": 0.01922607421875,
+      "learning_rate": 1e-06,
+      "loss": 0.0423,
+      "num_tokens": 2517637.0,
+      "reward": 0.1875,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.1875,
+      "step": 125
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 492.96875,
+      "epoch": 0.24766584766584768,
+      "grad_norm": 0.8350574297158757,
+      "kl": 0.023712158203125,
+      "learning_rate": 1e-06,
+      "loss": 0.0469,
+      "num_tokens": 2538338.0,
+      "reward": 0.34375,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.34375,
+      "step": 126
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 573.5,
+      "epoch": 0.24963144963144962,
+      "grad_norm": 0.8216738883653201,
+      "kl": 0.02204132080078125,
+      "learning_rate": 1e-06,
+      "loss": 0.0218,
+      "num_tokens": 2560112.0,
+      "reward": 0.28125,
+      "reward_std": 0.30935921147465706,
+      "rewards/reward_func": 0.28125,
+      "step": 127
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 529.3125,
+      "epoch": 0.2515970515970516,
+      "grad_norm": 0.78248285660239,
+      "kl": 0.02326202392578125,
+      "learning_rate": 1e-06,
+      "loss": -0.044,
+      "num_tokens": 2580546.0,
+      "reward": 0.34375,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.34375,
+      "step": 128
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 602.1875,
+      "epoch": 0.25356265356265356,
+      "grad_norm": 0.2772418236153871,
+      "kl": 0.019317626953125,
+      "learning_rate": 1e-06,
+      "loss": 0.0027,
+      "num_tokens": 2604460.0,
+      "reward": 0.15625,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.15625,
+      "step": 129
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 548.0625,
+      "epoch": 0.25552825552825553,
+      "grad_norm": 0.6625525706905956,
+      "kl": 0.02095794677734375,
+      "learning_rate": 1e-06,
+      "loss": 0.0115,
+      "num_tokens": 2624972.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 130
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 585.9375,
+      "epoch": 0.2574938574938575,
+      "grad_norm": 0.752169210277677,
+      "kl": 0.021514892578125,
+      "learning_rate": 1e-06,
+      "loss": 0.0359,
+      "num_tokens": 2648150.0,
+      "reward": 0.375,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.375,
+      "step": 131
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 574.0625,
+      "epoch": 0.2594594594594595,
+      "grad_norm": 0.5002363204112915,
+      "kl": 0.0212249755859375,
+      "learning_rate": 1e-06,
+      "loss": 0.023,
+      "num_tokens": 2670338.0,
+      "reward": 0.28125,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.28125,
+      "step": 132
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 526.625,
+      "epoch": 0.26142506142506144,
+      "grad_norm": 0.9674758127355131,
+      "kl": 0.0261077880859375,
+      "learning_rate": 1e-06,
+      "loss": 0.0159,
+      "num_tokens": 2691642.0,
+      "reward": 0.34375,
+      "reward_std": 0.30935921147465706,
+      "rewards/reward_func": 0.34375,
+      "step": 133
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 412.59375,
+      "epoch": 0.2633906633906634,
+      "grad_norm": 1.0324181511597994,
+      "kl": 0.0365447998046875,
+      "learning_rate": 1e-06,
+      "loss": 0.0109,
+      "num_tokens": 2708471.0,
+      "reward": 0.53125,
+      "reward_std": 0.30935921147465706,
+      "rewards/reward_func": 0.53125,
+      "step": 134
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 501.1875,
+      "epoch": 0.26535626535626533,
+      "grad_norm": 0.646085784140391,
+      "kl": 0.02829742431640625,
+      "learning_rate": 1e-06,
+      "loss": -0.0245,
+      "num_tokens": 2728631.0,
+      "reward": 0.34375,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.34375,
+      "step": 135
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 524.59375,
+      "epoch": 0.2673218673218673,
+      "grad_norm": 0.5827914647557386,
+      "kl": 0.0233001708984375,
+      "learning_rate": 1e-06,
+      "loss": -0.0133,
+      "num_tokens": 2748682.0,
+      "reward": 0.375,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.375,
+      "step": 136
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 493.625,
+      "epoch": 0.26928746928746927,
+      "grad_norm": 0.5383283228577356,
+      "kl": 0.02911376953125,
+      "learning_rate": 1e-06,
+      "loss": 0.006,
+      "num_tokens": 2768384.0,
+      "reward": 0.15625,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.15625,
+      "step": 137
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 563.34375,
+      "epoch": 0.27125307125307124,
+      "grad_norm": 0.44514691751088103,
+      "kl": 0.02154541015625,
+      "learning_rate": 1e-06,
+      "loss": -0.045,
+      "num_tokens": 2790263.0,
+      "reward": 0.15625,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.15625,
+      "step": 138
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 520.8125,
+      "epoch": 0.2732186732186732,
+      "grad_norm": 0.7190292298516152,
+      "kl": 0.0236663818359375,
+      "learning_rate": 1e-06,
+      "loss": 0.0132,
+      "num_tokens": 2810351.0,
+      "reward": 0.375,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.375,
+      "step": 139
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 527.34375,
+      "epoch": 0.2751842751842752,
+      "grad_norm": 0.7242965068485637,
+      "kl": 0.0230255126953125,
+      "learning_rate": 1e-06,
+      "loss": 0.001,
+      "num_tokens": 2830716.0,
+      "reward": 0.34375,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.34375,
+      "step": 140
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 565.8125,
+      "epoch": 0.27714987714987716,
+      "grad_norm": 0.6228089750555335,
+      "kl": 0.02398681640625,
+      "learning_rate": 1e-06,
+      "loss": 0.0041,
+      "num_tokens": 2853810.0,
+      "reward": 0.34375,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.34375,
+      "step": 141
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 555.1875,
+      "epoch": 0.2791154791154791,
+      "grad_norm": 0.37035042072698693,
+      "kl": 0.02410888671875,
+      "learning_rate": 1e-06,
+      "loss": 0.0047,
+      "num_tokens": 2875056.0,
+      "reward": 0.3125,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.3125,
+      "step": 142
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 589.65625,
+      "epoch": 0.2810810810810811,
+      "grad_norm": 0.658035150283911,
+      "kl": 0.02394866943359375,
+      "learning_rate": 1e-06,
+      "loss": 0.025,
+      "num_tokens": 2897357.0,
+      "reward": 0.4375,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.4375,
+      "step": 143
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 543.8125,
+      "epoch": 0.28304668304668307,
+      "grad_norm": 0.7501331474001123,
+      "kl": 0.025238037109375,
+      "learning_rate": 1e-06,
+      "loss": -0.0123,
+      "num_tokens": 2917753.0,
+      "reward": 0.25,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.25,
+      "step": 144
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 612.28125,
+      "epoch": 0.28501228501228504,
+      "grad_norm": 0.5446469883013957,
+      "kl": 0.0233612060546875,
+      "learning_rate": 1e-06,
+      "loss": 0.0054,
+      "num_tokens": 2940774.0,
+      "reward": 0.3125,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.3125,
+      "step": 145
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 539.78125,
+      "epoch": 0.28697788697788695,
+      "grad_norm": 0.6092756185830317,
+      "kl": 0.02280426025390625,
+      "learning_rate": 1e-06,
+      "loss": 0.0263,
+      "num_tokens": 2961247.0,
+      "reward": 0.34375,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.34375,
+      "step": 146
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 466.28125,
+      "epoch": 0.2889434889434889,
+      "grad_norm": 0.6941889736363939,
+      "kl": 0.0313873291015625,
+      "learning_rate": 1e-06,
+      "loss": -0.0046,
+      "num_tokens": 2979622.0,
+      "reward": 0.09375,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.09375,
+      "step": 147
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 567.53125,
+      "epoch": 0.2909090909090909,
+      "grad_norm": 0.3736888500181184,
+      "kl": 0.023223876953125,
+      "learning_rate": 1e-06,
+      "loss": 0.0099,
+      "num_tokens": 3002039.0,
+      "reward": 0.15625,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.15625,
+      "step": 148
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 467.59375,
+      "epoch": 0.29287469287469287,
+      "grad_norm": 0.8319282300423898,
+      "kl": 0.02935791015625,
+      "learning_rate": 1e-06,
+      "loss": 0.043,
+      "num_tokens": 3019816.0,
+      "reward": 0.375,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.375,
+      "step": 149
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 492.25,
+      "epoch": 0.29484029484029484,
+      "grad_norm": 0.7325276211486558,
+      "kl": 0.0263671875,
+      "learning_rate": 1e-06,
+      "loss": 0.0293,
+      "num_tokens": 3039146.0,
+      "reward": 0.46875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.46875,
+      "step": 150
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 592.1875,
+      "epoch": 0.2968058968058968,
+      "grad_norm": 0.43633918201867017,
+      "kl": 0.0228424072265625,
+      "learning_rate": 1e-06,
+      "loss": -0.0122,
+      "num_tokens": 3062102.0,
+      "reward": 0.34375,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.34375,
+      "step": 151
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 563.1875,
+      "epoch": 0.2987714987714988,
+      "grad_norm": 0.38992921040393713,
+      "kl": 0.025421142578125,
+      "learning_rate": 1e-06,
+      "loss": 0.0327,
+      "num_tokens": 3084284.0,
+      "reward": 0.28125,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.28125,
+      "step": 152
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 634.9375,
+      "epoch": 0.30073710073710075,
+      "grad_norm": 0.41943188112426816,
+      "kl": 0.0197906494140625,
+      "learning_rate": 1e-06,
+      "loss": 0.001,
+      "num_tokens": 3108216.0,
+      "reward": 0.1875,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.1875,
+      "step": 153
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 636.9375,
+      "epoch": 0.3027027027027027,
+      "grad_norm": 0.19968498604663434,
+      "kl": 0.02194976806640625,
+      "learning_rate": 1e-06,
+      "loss": 0.0001,
+      "num_tokens": 3133100.0,
+      "reward": 0.09375,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.09375,
+      "step": 154
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 568.21875,
+      "epoch": 0.3046683046683047,
+      "grad_norm": 0.7545074739064522,
+      "kl": 0.02685546875,
+      "learning_rate": 1e-06,
+      "loss": 0.04,
+      "num_tokens": 3154579.0,
+      "reward": 0.46875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.46875,
+      "step": 155
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 529.8125,
+      "epoch": 0.30663390663390666,
+      "grad_norm": 0.6884386708248261,
+      "kl": 0.025360107421875,
+      "learning_rate": 1e-06,
+      "loss": -0.0075,
+      "num_tokens": 3174845.0,
+      "reward": 0.40625,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.40625,
+      "step": 156
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 513.0625,
+      "epoch": 0.3085995085995086,
+      "grad_norm": 0.5265944726858988,
+      "kl": 0.0264892578125,
+      "learning_rate": 1e-06,
+      "loss": -0.0245,
+      "num_tokens": 3195631.0,
+      "reward": 0.21875,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.21875,
+      "step": 157
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 520.0,
+      "epoch": 0.31056511056511055,
+      "grad_norm": 0.45924555811174794,
+      "kl": 0.026092529296875,
+      "learning_rate": 1e-06,
+      "loss": -0.0189,
+      "num_tokens": 3216143.0,
+      "reward": 0.21875,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.21875,
+      "step": 158
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 568.15625,
+      "epoch": 0.3125307125307125,
+      "grad_norm": 0.5301940769841443,
+      "kl": 0.025177001953125,
+      "learning_rate": 1e-06,
+      "loss": -0.0008,
+      "num_tokens": 3238264.0,
+      "reward": 0.15625,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.15625,
+      "step": 159
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 508.59375,
+      "epoch": 0.3144963144963145,
+      "grad_norm": 0.8755698562076372,
+      "kl": 0.0290069580078125,
+      "learning_rate": 1e-06,
+      "loss": 0.0536,
+      "num_tokens": 3257837.0,
+      "reward": 0.34375,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.34375,
+      "step": 160
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 481.15625,
+      "epoch": 0.31646191646191646,
+      "grad_norm": 0.6081291396581243,
+      "kl": 0.0274810791015625,
+      "learning_rate": 1e-06,
+      "loss": 0.0018,
+      "num_tokens": 3276820.0,
+      "reward": 0.4375,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.4375,
+      "step": 161
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 598.28125,
+      "epoch": 0.31842751842751843,
+      "grad_norm": 0.5215378127068343,
+      "kl": 0.0220794677734375,
+      "learning_rate": 1e-06,
+      "loss": 0.0012,
+      "num_tokens": 3299357.0,
+      "reward": 0.375,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.375,
+      "step": 162
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 487.09375,
+      "epoch": 0.3203931203931204,
+      "grad_norm": 0.5432850798950747,
+      "kl": 0.03083038330078125,
+      "learning_rate": 1e-06,
+      "loss": -0.0034,
+      "num_tokens": 3318084.0,
+      "reward": 0.5,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.5,
+      "step": 163
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 531.125,
+      "epoch": 0.32235872235872237,
+      "grad_norm": 0.46027668679847555,
+      "kl": 0.026092529296875,
+      "learning_rate": 1e-06,
+      "loss": 0.0154,
+      "num_tokens": 3338328.0,
+      "reward": 0.125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.125,
+      "step": 164
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 624.8125,
+      "epoch": 0.32432432432432434,
+      "grad_norm": 0.18474075838750667,
+      "kl": 0.02459716796875,
+      "learning_rate": 1e-06,
+      "loss": -0.0011,
+      "num_tokens": 3361788.0,
+      "reward": 0.21875,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.21875,
+      "step": 165
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 549.78125,
+      "epoch": 0.3262899262899263,
+      "grad_norm": 0.7718886882448309,
+      "kl": 0.027069091796875,
+      "learning_rate": 1e-06,
+      "loss": -0.005,
+      "num_tokens": 3382981.0,
+      "reward": 0.46875,
+      "reward_std": 0.30935921147465706,
+      "rewards/reward_func": 0.46875,
+      "step": 166
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 575.78125,
+      "epoch": 0.32825552825552823,
+      "grad_norm": 0.5895346584970902,
+      "kl": 0.02323150634765625,
+      "learning_rate": 1e-06,
+      "loss": -0.0178,
+      "num_tokens": 3405274.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 167
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 576.875,
+      "epoch": 0.3302211302211302,
+      "grad_norm": 0.4959807619539926,
+      "kl": 0.0241851806640625,
+      "learning_rate": 1e-06,
+      "loss": 0.0022,
+      "num_tokens": 3427392.0,
+      "reward": 0.21875,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.21875,
+      "step": 168
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 561.78125,
+      "epoch": 0.33218673218673217,
+      "grad_norm": 0.8681759456851222,
+      "kl": 0.0249176025390625,
+      "learning_rate": 1e-06,
+      "loss": 0.0274,
+      "num_tokens": 3448681.0,
+      "reward": 0.34375,
+      "reward_std": 0.30935921147465706,
+      "rewards/reward_func": 0.34375,
+      "step": 169
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 483.53125,
+      "epoch": 0.33415233415233414,
+      "grad_norm": 0.5297566985301704,
+      "kl": 0.0336761474609375,
+      "learning_rate": 1e-06,
+      "loss": -0.0019,
+      "num_tokens": 3467212.0,
+      "reward": 0.21875,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.21875,
+      "step": 170
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 581.40625,
+      "epoch": 0.3361179361179361,
+      "grad_norm": 0.6192719333892511,
+      "kl": 0.025787353515625,
+      "learning_rate": 1e-06,
+      "loss": 0.0075,
+      "num_tokens": 3489355.0,
+      "reward": 0.25,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.25,
+      "step": 171
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 709.09375,
+      "epoch": 0.3380835380835381,
+      "grad_norm": 0.47423370876978443,
+      "kl": 0.0204620361328125,
+      "learning_rate": 1e-06,
+      "loss": -0.0251,
+      "num_tokens": 3515834.0,
+      "reward": 0.21875,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.21875,
+      "step": 172
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 614.21875,
+      "epoch": 0.34004914004914005,
+      "grad_norm": 0.6479887044288859,
+      "kl": 0.0229034423828125,
+      "learning_rate": 1e-06,
+      "loss": -0.0212,
+      "num_tokens": 3539067.0,
+      "reward": 0.3125,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.3125,
+      "step": 173
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 552.6875,
+      "epoch": 0.342014742014742,
+      "grad_norm": 0.4712074892004097,
+      "kl": 0.0272216796875,
+      "learning_rate": 1e-06,
+      "loss": -0.0209,
+      "num_tokens": 3560435.0,
+      "reward": 0.21875,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.21875,
+      "step": 174
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 599.75,
+      "epoch": 0.343980343980344,
+      "grad_norm": 0.7325904251977917,
+      "kl": 0.0259552001953125,
+      "learning_rate": 1e-06,
+      "loss": 0.0087,
+      "num_tokens": 3582617.0,
+      "reward": 0.25,
+      "reward_std": 0.3535533845424652,
+      "rewards/reward_func": 0.25,
+      "step": 175
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 589.46875,
+      "epoch": 0.34594594594594597,
+      "grad_norm": 0.718404368627785,
+      "kl": 0.024017333984375,
+      "learning_rate": 1e-06,
+      "loss": 0.0069,
+      "num_tokens": 3606486.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 176
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 525.0625,
+      "epoch": 0.34791154791154794,
+      "grad_norm": 0.5807447821668908,
+      "kl": 0.0262451171875,
+      "learning_rate": 1e-06,
+      "loss": 0.006,
+      "num_tokens": 3626528.0,
+      "reward": 0.3125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.3125,
+      "step": 177
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 539.875,
+      "epoch": 0.34987714987714985,
+      "grad_norm": 0.3836295461022733,
+      "kl": 0.025665283203125,
+      "learning_rate": 1e-06,
+      "loss": -0.0233,
+      "num_tokens": 3647890.0,
+      "reward": 0.25,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.25,
+      "step": 178
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 641.28125,
+      "epoch": 0.3518427518427518,
+      "grad_norm": 0.3325268331032711,
+      "kl": 0.02361297607421875,
+      "learning_rate": 1e-06,
+      "loss": 0.0117,
+      "num_tokens": 3672977.0,
+      "reward": 0.1875,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.1875,
+      "step": 179
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 611.1875,
+      "epoch": 0.3538083538083538,
+      "grad_norm": 0.5612204711943589,
+      "kl": 0.025909423828125,
+      "learning_rate": 1e-06,
+      "loss": 0.014,
+      "num_tokens": 3696201.0,
+      "reward": 0.28125,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.28125,
+      "step": 180
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 587.125,
+      "epoch": 0.35577395577395576,
+      "grad_norm": 0.48905557864261767,
+      "kl": 0.0284881591796875,
+      "learning_rate": 1e-06,
+      "loss": -0.0135,
+      "num_tokens": 3718645.0,
+      "reward": 0.21875,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.21875,
+      "step": 181
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 532.8125,
+      "epoch": 0.35773955773955773,
+      "grad_norm": 0.6196793685491477,
+      "kl": 0.0283050537109375,
+      "learning_rate": 1e-06,
+      "loss": -0.021,
+      "num_tokens": 3739419.0,
+      "reward": 0.34375,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.34375,
+      "step": 182
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 491.25,
+      "epoch": 0.3597051597051597,
+      "grad_norm": 0.7943049071356907,
+      "kl": 0.0270843505859375,
+      "learning_rate": 1e-06,
+      "loss": 0.0343,
+      "num_tokens": 3758175.0,
+      "reward": 0.3125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.3125,
+      "step": 183
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 629.3125,
+      "epoch": 0.3616707616707617,
+      "grad_norm": 0.4358875914582024,
+      "kl": 0.02367401123046875,
+      "learning_rate": 1e-06,
+      "loss": 0.0133,
+      "num_tokens": 3782789.0,
+      "reward": 0.21875,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.21875,
+      "step": 184
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 540.59375,
+      "epoch": 0.36363636363636365,
+      "grad_norm": 0.7097269605040931,
+      "kl": 0.028656005859375,
+      "learning_rate": 1e-06,
+      "loss": 0.023,
+      "num_tokens": 3804202.0,
+      "reward": 0.25,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.25,
+      "step": 185
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 517.40625,
+      "epoch": 0.3656019656019656,
+      "grad_norm": 0.8226505283368968,
+      "kl": 0.0278778076171875,
+      "learning_rate": 1e-06,
+      "loss": -0.0374,
+      "num_tokens": 3824813.0,
+      "reward": 0.3125,
+      "reward_std": 0.3535533845424652,
+      "rewards/reward_func": 0.3125,
+      "step": 186
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 499.3125,
+      "epoch": 0.3675675675675676,
+      "grad_norm": 0.8593511870826509,
+      "kl": 0.026092529296875,
+      "learning_rate": 1e-06,
+      "loss": 0.0443,
+      "num_tokens": 3844419.0,
+      "reward": 0.3125,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.3125,
+      "step": 187
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 497.5,
+      "epoch": 0.36953316953316956,
+      "grad_norm": 0.6026343875240253,
+      "kl": 0.0312042236328125,
+      "learning_rate": 1e-06,
+      "loss": -0.0226,
+      "num_tokens": 3863979.0,
+      "reward": 0.15625,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.15625,
+      "step": 188
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 547.0625,
+      "epoch": 0.3714987714987715,
+      "grad_norm": 0.6577878892956752,
+      "kl": 0.030914306640625,
+      "learning_rate": 1e-06,
+      "loss": -0.002,
+      "num_tokens": 3885407.0,
+      "reward": 0.21875,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.21875,
+      "step": 189
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 605.0625,
+      "epoch": 0.37346437346437344,
+      "grad_norm": 0.45127850046328344,
+      "kl": 0.024169921875,
+      "learning_rate": 1e-06,
+      "loss": -0.003,
+      "num_tokens": 3908713.0,
+      "reward": 0.40625,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.40625,
+      "step": 190
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 529.46875,
+      "epoch": 0.3754299754299754,
+      "grad_norm": 0.5469774623716551,
+      "kl": 0.025115966796875,
+      "learning_rate": 1e-06,
+      "loss": -0.0056,
+      "num_tokens": 3928856.0,
+      "reward": 0.28125,
+      "reward_std": 0.13258251920342445,
+      "rewards/reward_func": 0.28125,
+      "step": 191
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 475.15625,
+      "epoch": 0.3773955773955774,
+      "grad_norm": 0.9374126682104605,
+      "kl": 0.0326385498046875,
+      "learning_rate": 1e-06,
+      "loss": 0.0543,
+      "num_tokens": 3947241.0,
+      "reward": 0.3125,
+      "reward_std": 0.2651650384068489,
+      "rewards/reward_func": 0.3125,
+      "step": 192
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 547.1875,
+      "epoch": 0.37936117936117936,
+      "grad_norm": 0.5527175997380437,
+      "kl": 0.0286407470703125,
+      "learning_rate": 1e-06,
+      "loss": 0.0571,
+      "num_tokens": 3968765.0,
+      "reward": 0.3125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.3125,
+      "step": 193
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 581.53125,
+      "epoch": 0.38132678132678133,
+      "grad_norm": 0.5091715028558698,
+      "kl": 0.02716064453125,
+      "learning_rate": 1e-06,
+      "loss": -0.0032,
+      "num_tokens": 3990744.0,
+      "reward": 0.4375,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.4375,
+      "step": 194
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 545.3125,
+      "epoch": 0.3832923832923833,
+      "grad_norm": 0.6906699369220881,
+      "kl": 0.026611328125,
+      "learning_rate": 1e-06,
+      "loss": 0.0317,
+      "num_tokens": 4011242.0,
+      "reward": 0.28125,
+      "reward_std": 0.22097086533904076,
+      "rewards/reward_func": 0.28125,
+      "step": 195
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 553.0625,
+      "epoch": 0.38525798525798527,
+      "grad_norm": 0.803923186478945,
+      "kl": 0.026153564453125,
+      "learning_rate": 1e-06,
+      "loss": 0.0114,
+      "num_tokens": 4032454.0,
+      "reward": 0.34375,
+      "reward_std": 0.30935921147465706,
+      "rewards/reward_func": 0.34375,
+      "step": 196
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 514.875,
+      "epoch": 0.38722358722358724,
+      "grad_norm": 0.5425045400963313,
+      "kl": 0.029937744140625,
+      "learning_rate": 1e-06,
+      "loss": -0.0032,
+      "num_tokens": 4053082.0,
+      "reward": 0.3125,
+      "reward_std": 0.0883883461356163,
+      "rewards/reward_func": 0.3125,
+      "step": 197
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 677.78125,
+      "epoch": 0.3891891891891892,
+      "grad_norm": 0.6672952970651821,
+      "kl": 0.0213623046875,
+      "learning_rate": 1e-06,
+      "loss": 0.0023,
+      "num_tokens": 4078691.0,
+      "reward": 0.125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.125,
+      "step": 198
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 512.03125,
+      "epoch": 0.3911547911547912,
+      "grad_norm": 0.476359304779961,
+      "kl": 0.035552978515625,
+      "learning_rate": 1e-06,
+      "loss": 0.0066,
+      "num_tokens": 4099720.0,
+      "reward": 0.21875,
+      "reward_std": 0.04419417306780815,
+      "rewards/reward_func": 0.21875,
+      "step": 199
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 743.375,
+      "epoch": 0.3931203931203931,
+      "grad_norm": 0.4679126082574762,
+      "kl": 0.0198974609375,
+      "learning_rate": 1e-06,
+      "loss": -0.0029,
+      "num_tokens": 4127002.0,
+      "reward": 0.3125,
+      "reward_std": 0.1767766922712326,
+      "rewards/reward_func": 0.3125,
+      "step": 200
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 508,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-200/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd8a999b49f0161be12fb454c7767dab935535c70c583f510c6a2dc6c7bfe2d7
+size 7544