{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9884169884169884,
  "eval_steps": 100,
  "global_step": 32,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 291.36697788238524,
      "epoch": 0.15444015444015444,
      "grad_norm": 91.9554297295678,
      "kl": 0.3540959358215332,
      "learning_rate": 1.9937122098932428e-05,
      "loss": 0.0142,
      "reward": 0.6166573823895305,
      "reward_std": 0.452694922266528,
      "rewards/accuracy_reward": 0.10803571995347738,
      "rewards/cosine_scaled_reward": -0.10179503666004167,
      "rewards/format_reward": 0.5000000233761966,
      "rewards/reasoning_steps_reward": 0.110416672937572,
      "step": 5
    },
    {
      "completion_length": 43.16875188350677,
      "epoch": 0.3088803088803089,
      "grad_norm": 3.33245795174105,
      "kl": 0.8302490234375,
      "learning_rate": 1.78183148246803e-05,
      "loss": 0.0332,
      "reward": 0.9400907799601554,
      "reward_std": 0.19169574869779354,
      "rewards/accuracy_reward": 0.024107144121080636,
      "rewards/cosine_scaled_reward": 0.0037812213704455644,
      "rewards/format_reward": 0.9107143148779869,
      "rewards/reasoning_steps_reward": 0.0014880953822284937,
      "step": 10
    },
    {
      "completion_length": 15.207143580913543,
      "epoch": 0.46332046332046334,
      "grad_norm": 0.8155119970711466,
      "kl": 1.6966796875,
      "learning_rate": 1.3302790619551673e-05,
      "loss": 0.0679,
      "reward": 0.9947075441479682,
      "reward_std": 0.005296425729534348,
      "rewards/accuracy_reward": 0.0,
      "rewards/cosine_scaled_reward": -0.0017210765145136975,
      "rewards/format_reward": 0.9964285731315613,
      "rewards/reasoning_steps_reward": 0.0,
      "step": 15
    },
    {
      "completion_length": 14.96160796880722,
      "epoch": 0.6177606177606177,
      "grad_norm": 0.5795656668267815,
      "kl": 1.59228515625,
      "learning_rate": 7.774790660436857e-06,
      "loss": 0.0637,
      "reward": 0.9939029954373837,
      "reward_std": 0.006596084152567983,
      "rewards/accuracy_reward": 0.0,
      "rewards/cosine_scaled_reward": -0.0016327791643561795,
      "rewards/format_reward": 0.9946428596973419,
      "rewards/reasoning_steps_reward": 0.0008928571827709675,
      "step": 20
    },
    {
      "completion_length": 14.393750703334808,
      "epoch": 0.7722007722007722,
      "grad_norm": 4.257970780643175,
      "kl": 1.46865234375,
      "learning_rate": 2.9289321881345257e-06,
      "loss": 0.0588,
      "reward": 0.989137577265501,
      "reward_std": 0.013341282614055672,
      "rewards/accuracy_reward": 0.0,
      "rewards/cosine_scaled_reward": -0.0019339090387802571,
      "rewards/format_reward": 0.9910714328289032,
      "rewards/reasoning_steps_reward": 0.0,
      "step": 25
    },
    {
      "completion_length": 13.3160719871521,
      "epoch": 0.9266409266409267,
      "grad_norm": 0.5366189541461951,
      "kl": 1.44560546875,
      "learning_rate": 2.507208781817638e-07,
      "loss": 0.0578,
      "reward": 0.9885662972927094,
      "reward_std": 0.014164349375778328,
      "rewards/accuracy_reward": 0.0,
      "rewards/cosine_scaled_reward": -0.0016123341440106743,
      "rewards/format_reward": 0.9901785761117935,
      "rewards/reasoning_steps_reward": 0.0,
      "step": 30
    },
    {
      "completion_length": 12.924107730388641,
      "epoch": 0.9884169884169884,
      "kl": 1.577392578125,
      "reward": 0.9940091818571091,
      "reward_std": 0.006472988553753112,
      "rewards/accuracy_reward": 0.0,
      "rewards/cosine_scaled_reward": -0.0015265825350070372,
      "rewards/format_reward": 0.9955357164144516,
      "rewards/reasoning_steps_reward": 0.0,
      "step": 32,
      "total_flos": 0.0,
      "train_loss": 0.049734286265447736,
      "train_runtime": 1467.0643,
      "train_samples_per_second": 2.469,
      "train_steps_per_second": 0.022
    }
  ],
  "logging_steps": 5,
  "max_steps": 32,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}