{
    "epoch": 1.0,
    "eval_logits/chosen": -2.861914873123169,
    "eval_logits/rejected": -2.5182931423187256,
    "eval_logps/chosen": -188.0933074951172,
    "eval_logps/rejected": -167.61656188964844,
    "eval_loss": 0.6934889554977417,
    "eval_rewards/accuracies": 0.5833333134651184,
    "eval_rewards/chosen": -0.024234572425484657,
    "eval_rewards/margins": 0.004475079942494631,
    "eval_rewards/rejected": -0.0287096556276083,
    "eval_runtime": 17.5046,
    "eval_samples": 163,
    "eval_samples_per_second": 9.312,
    "eval_steps_per_second": 0.343,
    "train_loss": 0.6464094665815245,
    "train_runtime": 1236.0981,
    "train_samples": 6754,
    "train_samples_per_second": 5.464,
    "train_steps_per_second": 0.086
}