{ "epoch": 2.999098751126561, "eval_kl/chosen": 0.00018642864597495645, "eval_kl/rejected": 0.0005706130177713931, "eval_logits/chosen": -1.5, "eval_logits/rejected": -1.46875, "eval_logps/chosen": -322.0, "eval_logps/rejected": -300.0, "eval_loss": 0.6433281302452087, "eval_nll_loss": 1.0546875, "eval_ref_probs/chosen": 0.0038271306548267603, "eval_ref_probs/rejected": 0.0034569373819977045, "eval_rejected_term/max": 0.0005706130177713931, "eval_rejected_term/min": 0.0005706130177713931, "eval_rewards/accuracies": 0.7360000014305115, "eval_rewards/chosen": -0.025146484375, "eval_rewards/margins": 0.2412109375, "eval_rewards/rejected": -0.265625, "eval_runtime": 29.8044, "eval_samples_per_second": 33.552, "eval_steps_per_second": 8.388 }