eval_results.json

{
    "epoch": 1.0,
    "eval_logits/chosen": -0.4606437087059021,
    "eval_logits/rejected": 0.6826901435852051,
    "eval_logps/chosen": -477.88958740234375,
    "eval_logps/rejected": -563.9073486328125,
    "eval_loss": 0.48932957649230957,
    "eval_rewards/accuracies": 0.77182537317276,
    "eval_rewards/chosen": -1.9378634691238403,
    "eval_rewards/margins": 1.0834674835205078,
    "eval_rewards/rejected": -3.0213310718536377,
    "eval_runtime": 279.233,
    "eval_samples": 2000,
    "eval_samples_per_second": 7.162,
    "eval_steps_per_second": 0.226
}