{
    "epoch": 0.9989071038251366,
    "eval_logits/chosen": -1.3906669616699219,
    "eval_logits/rejected": -1.3494510650634766,
    "eval_logps/chosen": -0.5431549549102783,
    "eval_logps/rejected": -0.7273014187812805,
    "eval_loss": 1.8781323432922363,
    "eval_rewards/accuracies": 0.7710843086242676,
    "eval_rewards/chosen": -5.431549549102783,
    "eval_rewards/margins": 1.841464638710022,
    "eval_rewards/rejected": -7.273014068603516,
    "eval_runtime": 33.4862,
    "eval_samples": 1318,
    "eval_samples_per_second": 39.359,
    "eval_steps_per_second": 2.479
}