{
    "epoch": 1.0,
    "eval_logits/chosen": -1.238540530204773,
    "eval_logits/rejected": -1.3014811277389526,
    "eval_logps/chosen": -481.2091979980469,
    "eval_logps/rejected": -755.0648193359375,
    "eval_loss": 0.4001818299293518,
    "eval_rewards/accuracies": 0.8446428775787354,
    "eval_rewards/chosen": -2.249953508377075,
    "eval_rewards/margins": 2.8106350898742676,
    "eval_rewards/rejected": -5.060588836669922,
    "eval_runtime": 193.9304,
    "eval_samples": 4461,
    "eval_samples_per_second": 23.003,
    "eval_steps_per_second": 0.722
}