{
    "epoch": 1.99,
    "eval_logits/chosen": -2.5726864337921143,
    "eval_logits/rejected": -2.556718587875366,
    "eval_logps/chosen": -85.11767578125,
    "eval_logps/rejected": -133.85464477539062,
    "eval_loss": 0.12992651760578156,
    "eval_rewards/accuracies": 0.9520547986030579,
    "eval_rewards/chosen": -2.199751138687134,
    "eval_rewards/margins": 4.689663887023926,
    "eval_rewards/rejected": -6.889414310455322,
    "eval_runtime": 288.3554,
    "eval_samples": 876,
    "eval_samples_per_second": 3.038,
    "eval_steps_per_second": 3.038,
    "train_loss": 0.20939183043777396,
    "train_runtime": 5162.9911,
    "train_samples": 3505,
    "train_samples_per_second": 1.358,
    "train_steps_per_second": 0.042
}