{
    "epoch": 1.0,
    "eval_logits/chosen": -1.2245557308197021,
    "eval_logits/rejected": 1.1489198207855225,
    "eval_logps/chosen": -197.46658325195312,
    "eval_logps/rejected": -414.81317138671875,
    "eval_loss": 0.4863126873970032,
    "eval_rewards/accuracies": 0.77734375,
    "eval_rewards/chosen": -0.45359015464782715,
    "eval_rewards/margins": 1.201076626777649,
    "eval_rewards/rejected": -1.6546669006347656,
    "eval_runtime": 103.3892,
    "eval_samples": 1000,
    "eval_samples_per_second": 9.672,
    "eval_steps_per_second": 0.31
}