{
    "epoch": 3.0,
    "eval_logits/chosen": -2.198760986328125,
    "eval_logits/rejected": -2.1614327430725098,
    "eval_logps/chosen": -278.5389709472656,
    "eval_logps/rejected": -310.5012512207031,
    "eval_loss": 0.6269853711128235,
    "eval_rewards/accuracies": 0.6924999952316284,
    "eval_rewards/chosen": -7.6611433029174805,
    "eval_rewards/margins": 4.435902118682861,
    "eval_rewards/rejected": -12.0970458984375,
    "eval_runtime": 132.6261,
    "eval_samples": 3156,
    "eval_samples_per_second": 23.796,
    "eval_steps_per_second": 0.377,
    "train_loss": 0.24049862180692672,
    "train_runtime": 20948.1804,
    "train_samples": 82424,
    "train_samples_per_second": 9.519,
    "train_steps_per_second": 0.595
}