{
  "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": 66.10237426067714,
      "learning_rate": 4.545454545454545e-08,
      "logits/chosen": 117.53560638427734,
      "logits/rejected": 126.8960952758789,
      "logps/chosen": -335.40118408203125,
      "logps/rejected": -439.16552734375,
      "loss": 0.5,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.1895734597156398,
      "grad_norm": 65.8363279797723,
      "learning_rate": 4.545454545454545e-07,
      "logits/chosen": 135.0050811767578,
      "logits/rejected": 138.34999084472656,
      "logps/chosen": -396.04180908203125,
      "logps/rejected": -440.06195068359375,
      "loss": 0.4978,
      "rewards/accuracies": 0.4583333432674408,
      "rewards/chosen": -0.0022641660179942846,
      "rewards/margins": 0.03446006774902344,
      "rewards/rejected": -0.03672423213720322,
      "step": 10
    },
    {
      "epoch": 0.3791469194312796,
      "grad_norm": 30.020665129557617,
      "learning_rate": 4.885348141000122e-07,
      "logits/chosen": 123.14253234863281,
      "logits/rejected": 126.7535629272461,
      "logps/chosen": -354.2257995605469,
      "logps/rejected": -407.9169006347656,
      "loss": 0.4121,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.9093106985092163,
      "rewards/margins": 0.3055870234966278,
      "rewards/rejected": 0.6037237644195557,
      "step": 20
    },
    {
      "epoch": 0.5687203791469194,
      "grad_norm": 24.190229582222074,
      "learning_rate": 4.5025027361734613e-07,
      "logits/chosen": 145.62486267089844,
      "logits/rejected": 138.91897583007812,
      "logps/chosen": -376.835205078125,
      "logps/rejected": -432.81298828125,
      "loss": 0.3514,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.7815009355545044,
      "rewards/margins": 1.4220384359359741,
      "rewards/rejected": -0.640537440776825,
      "step": 30
    },
    {
      "epoch": 0.7582938388625592,
      "grad_norm": 24.559258380924415,
      "learning_rate": 3.893311157806091e-07,
      "logits/chosen": 134.8649139404297,
      "logits/rejected": 123.98270416259766,
      "logps/chosen": -324.2713317871094,
      "logps/rejected": -366.3738098144531,
      "loss": 0.3442,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 1.4961439371109009,
      "rewards/margins": 1.9210717678070068,
      "rewards/rejected": -0.42492780089378357,
      "step": 40
    },
    {
      "epoch": 0.9478672985781991,
      "grad_norm": 25.998652882986566,
      "learning_rate": 3.126631330646801e-07,
      "logits/chosen": 153.93051147460938,
      "logits/rejected": 157.34010314941406,
      "logps/chosen": -383.385009765625,
      "logps/rejected": -484.5369567871094,
      "loss": 0.3087,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 1.7879406213760376,
      "rewards/margins": 2.1984703540802,
      "rewards/rejected": -0.4105294644832611,
      "step": 50
    },
    {
      "epoch": 1.1374407582938388,
      "grad_norm": 20.98032293493931,
      "learning_rate": 2.2891223348923882e-07,
      "logits/chosen": 145.45932006835938,
      "logits/rejected": 149.6053466796875,
      "logps/chosen": -359.0306396484375,
      "logps/rejected": -456.782470703125,
      "loss": 0.257,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": 1.9487521648406982,
      "rewards/margins": 2.969599723815918,
      "rewards/rejected": -1.0208473205566406,
      "step": 60
    },
    {
      "epoch": 1.3270142180094786,
      "grad_norm": 19.65205364182723,
      "learning_rate": 1.4754491880085317e-07,
      "logits/chosen": 140.3638458251953,
      "logits/rejected": 141.1572265625,
      "logps/chosen": -328.19830322265625,
      "logps/rejected": -429.80487060546875,
      "loss": 0.2359,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 1.9472763538360596,
      "rewards/margins": 2.761829137802124,
      "rewards/rejected": -0.8145527839660645,
      "step": 70
    },
    {
      "epoch": 1.5165876777251186,
      "grad_norm": 19.33483383354362,
      "learning_rate": 7.775827023107834e-08,
      "logits/chosen": 127.70585632324219,
      "logits/rejected": 143.24069213867188,
      "logps/chosen": -311.0325622558594,
      "logps/rejected": -434.9661560058594,
      "loss": 0.2084,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": 2.130922794342041,
      "rewards/margins": 3.2173447608947754,
      "rewards/rejected": -1.0864222049713135,
      "step": 80
    },
    {
      "epoch": 1.7061611374407581,
      "grad_norm": 21.62644101836202,
      "learning_rate": 2.7440387297912122e-08,
      "logits/chosen": 128.0491180419922,
      "logits/rejected": 140.35018920898438,
      "logps/chosen": -341.66192626953125,
      "logps/rejected": -459.8998107910156,
      "loss": 0.2046,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": 2.238457202911377,
      "rewards/margins": 3.3493576049804688,
      "rewards/rejected": -1.110900640487671,
      "step": 90
    },
    {
      "epoch": 1.8957345971563981,
      "grad_norm": 19.399744221303617,
      "learning_rate": 2.27878296044029e-09,
      "logits/chosen": 134.3742218017578,
      "logits/rejected": 134.44503784179688,
      "logps/chosen": -334.2799377441406,
      "logps/rejected": -428.43621826171875,
      "loss": 0.1949,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": 2.384758472442627,
      "rewards/margins": 2.9683709144592285,
      "rewards/rejected": -0.5836124420166016,
      "step": 100
    },
    {
      "epoch": 1.971563981042654,
      "step": 104,
      "total_flos": 0.0,
      "train_loss": 0.29816230271871275,
      "train_runtime": 2165.9381,
      "train_samples_per_second": 6.233,
      "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
}