{ "epoch": 1.0, "eval_kl/n_epsilon_steps": 0.2517123222351074, "eval_kl/p_epsilon_steps": 0.7478595972061157, "eval_logits/chosen": -7.606959342956543, "eval_logits/rejected": -7.186291217803955, "eval_logps/chosen": -252.70166015625, "eval_logps/ref_chosen": -100.49356842041016, "eval_logps/ref_rejected": -94.06775665283203, "eval_logps/rejected": -314.7497253417969, "eval_loss": 0.6180921196937561, "eval_rewards/accuracies": 0.7577054500579834, "eval_rewards/chosen": -0.3851844370365143, "eval_rewards/margins": 0.17150068283081055, "eval_rewards/rejected": -0.5566850900650024, "eval_runtime": 48.3953, "eval_samples": 2339, "eval_samples_per_second": 48.331, "eval_steps_per_second": 1.529 }