{ "epoch": 90.0, "eval_kl/chosen": 4.1868162270475295e-07, "eval_kl/rejected": 5.305637262864593e-09, "eval_logits/chosen": -2.234375, "eval_logits/rejected": -2.234375, "eval_logps/chosen": -708.0, "eval_logps/rejected": -584.0, "eval_loss": 155.24598693847656, "eval_nll_loss": 2.5, "eval_ref_probs/chosen": 2.243688839407266e-21, "eval_ref_probs/rejected": 2.0227180643139018e-19, "eval_rejected_term/max": 5.305637262864593e-09, "eval_rejected_term/min": 5.305637262864593e-09, "eval_rewards/accuracies": 0.41200000047683716, "eval_rewards/chosen": -29.125, "eval_rewards/margins": -6.1875, "eval_rewards/rejected": -23.0, "eval_runtime": 25.2931, "eval_samples_per_second": 39.536, "eval_steps_per_second": 9.884 }