{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.971563981042654,
  "eval_steps": 128,
  "global_step": 104,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.018957345971563982,
      "grad_norm": 132.15363683823435,
      "learning_rate": 4.545454545454545e-08,
      "logits/chosen": 117.53560638427734,
      "logits/rejected": 126.8960952758789,
      "logps/chosen": -335.40118408203125,
      "logps/rejected": -439.16552734375,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.1895734597156398,
      "grad_norm": 131.13307610995435,
      "learning_rate": 4.545454545454545e-07,
      "logits/chosen": 135.01609802246094,
      "logits/rejected": 138.3571319580078,
      "logps/chosen": -395.9638366699219,
      "logps/rejected": -438.9075012207031,
      "loss": 0.7093,
      "rewards/accuracies": 0.4236111044883728,
      "rewards/chosen": 0.001634951215237379,
      "rewards/margins": -0.019367942586541176,
      "rewards/rejected": 0.021002890542149544,
      "step": 10
    },
    {
      "epoch": 0.3791469194312796,
      "grad_norm": 135.24638454105212,
      "learning_rate": 4.885348141000122e-07,
      "logits/chosen": 121.93367004394531,
      "logits/rejected": 125.64762115478516,
      "logps/chosen": -370.1551818847656,
      "logps/rejected": -423.16387939453125,
      "loss": 0.6449,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.11283999681472778,
      "rewards/margins": 0.27146396040916443,
      "rewards/rejected": -0.15862399339675903,
      "step": 20
    },
    {
      "epoch": 0.5687203791469194,
      "grad_norm": 116.62151753737555,
      "learning_rate": 4.5025027361734613e-07,
      "logits/chosen": 143.11940002441406,
      "logits/rejected": 136.62831115722656,
      "logps/chosen": -421.55322265625,
      "logps/rejected": -466.46875,
      "loss": 0.5776,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -1.4543992280960083,
      "rewards/margins": 0.8689262270927429,
      "rewards/rejected": -2.3233253955841064,
      "step": 30
    },
    {
      "epoch": 0.7582938388625592,
      "grad_norm": 105.72910301274642,
      "learning_rate": 3.893311157806091e-07,
      "logits/chosen": 126.14713287353516,
      "logits/rejected": 114.54765319824219,
      "logps/chosen": -404.1641540527344,
      "logps/rejected": -430.92340087890625,
      "loss": 0.541,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -2.4984982013702393,
      "rewards/margins": 1.1539076566696167,
      "rewards/rejected": -3.6524055004119873,
      "step": 40
    },
    {
      "epoch": 0.9478672985781991,
      "grad_norm": 121.8738070470713,
      "learning_rate": 3.126631330646801e-07,
      "logits/chosen": 141.29660034179688,
      "logits/rejected": 145.36276245117188,
      "logps/chosen": -465.7582092285156,
      "logps/rejected": -548.9434204101562,
      "loss": 0.5003,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -2.330719470977783,
      "rewards/margins": 1.3001314401626587,
      "rewards/rejected": -3.6308505535125732,
      "step": 50
    },
    {
      "epoch": 1.1374407582938388,
      "grad_norm": 57.66323824156206,
      "learning_rate": 2.2891223348923882e-07,
      "logits/chosen": 134.20419311523438,
      "logits/rejected": 137.76060485839844,
      "logps/chosen": -452.38360595703125,
      "logps/rejected": -538.2611083984375,
      "loss": 0.3088,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -2.7188973426818848,
      "rewards/margins": 2.375883102416992,
      "rewards/rejected": -5.094780445098877,
      "step": 60
    },
    {
      "epoch": 1.3270142180094786,
      "grad_norm": 43.75224088730381,
      "learning_rate": 1.4754491880085317e-07,
      "logits/chosen": 127.32780456542969,
      "logits/rejected": 129.4023895263672,
      "logps/chosen": -423.1866760253906,
      "logps/rejected": -524.3780517578125,
      "loss": 0.2043,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": -2.8021414279937744,
      "rewards/margins": 2.741067409515381,
      "rewards/rejected": -5.543208122253418,
      "step": 70
    },
    {
      "epoch": 1.5165876777251186,
      "grad_norm": 40.948501083863796,
      "learning_rate": 7.775827023107834e-08,
      "logits/chosen": 113.81268310546875,
      "logits/rejected": 130.70741271972656,
      "logps/chosen": -419.47991943359375,
      "logps/rejected": -537.3096923828125,
      "loss": 0.1753,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": -3.291445255279541,
      "rewards/margins": 2.9121522903442383,
      "rewards/rejected": -6.203597545623779,
      "step": 80
    },
    {
      "epoch": 1.7061611374407581,
      "grad_norm": 47.97569202986646,
      "learning_rate": 2.7440387297912122e-08,
      "logits/chosen": 112.64261627197266,
      "logits/rejected": 125.56978607177734,
      "logps/chosen": -451.9779357910156,
      "logps/rejected": -565.5565185546875,
      "loss": 0.163,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -3.2773444652557373,
      "rewards/margins": 3.116389513015747,
      "rewards/rejected": -6.393733978271484,
      "step": 90
    },
    {
      "epoch": 1.8957345971563981,
      "grad_norm": 54.70383764426309,
      "learning_rate": 2.27878296044029e-09,
      "logits/chosen": 118.79183197021484,
      "logits/rejected": 118.7574234008789,
      "logps/chosen": -441.4315490722656,
      "logps/rejected": -534.9552001953125,
      "loss": 0.1642,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -2.9728212356567383,
      "rewards/margins": 2.9367377758026123,
      "rewards/rejected": -5.9095587730407715,
      "step": 100
    },
    {
      "epoch": 1.971563981042654,
      "step": 104,
      "total_flos": 0.0,
      "train_loss": 0.3946822423201341,
      "train_runtime": 2172.3886,
      "train_samples_per_second": 6.214,
      "train_steps_per_second": 0.048
    }
  ],
  "logging_steps": 10,
  "max_steps": 104,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}