{
    "epoch": 3.0,
    "eval_logits/chosen": 1.3515625,
    "eval_logits/rejected": 2.953125,
    "eval_logps/chosen": -2048.0,
    "eval_logps/rejected": -2240.0,
    "eval_loss": 1.6630624532699585,
    "eval_rewards/accuracies": 0.603515625,
    "eval_rewards/chosen": -17.375,
    "eval_rewards/margins": 2.109375,
    "eval_rewards/rejected": -19.5,
    "eval_runtime": 46.7908,
    "eval_samples": 2000,
    "eval_samples_per_second": 42.743,
    "eval_steps_per_second": 0.684,
    "total_flos": 0.0,
    "train_loss": 0.22992089002752386,
    "train_runtime": 12100.6201,
    "train_samples": 61119,
    "train_samples_per_second": 15.153,
    "train_steps_per_second": 0.237
}