{
    "epoch": 1.0,
    "eval_logits/chosen": -2.515625,
    "eval_logits/rejected": -2.46875,
    "eval_logps/chosen": -388.0,
    "eval_logps/rejected": -488.0,
    "eval_loss": 0.4984453022480011,
    "eval_rewards/accuracies": 0.77734375,
    "eval_rewards/chosen": -0.9375,
    "eval_rewards/margins": 1.0078125,
    "eval_rewards/rejected": -1.9375,
    "eval_runtime": 40.2538,
    "eval_samples": 2000,
    "eval_samples_per_second": 49.685,
    "eval_steps_per_second": 0.795,
    "total_flos": 0.0,
    "train_loss": 0.0,
    "train_runtime": 1.4031,
    "train_samples": 100,
    "train_samples_per_second": 7126.865,
    "train_steps_per_second": 71.269
}