{ "epoch": 1.0, "eval_epsilon_dpo/beta": 0.00137312151491642, "eval_epsilon_dpo/beta_margin_grad_mean": -0.457266241312027, "eval_epsilon_dpo/beta_margin_grad_std": 0.07400882244110107, "eval_epsilon_dpo/beta_margin_mean": 0.17634743452072144, "eval_epsilon_dpo/beta_margin_std": 0.3081098198890686, "eval_epsilon_dpo/loss_margin_mean": 129.92251586914062, "eval_kl/n_epsilon_steps": 0.3146404027938843, "eval_kl/p_epsilon_steps": 0.6840753555297852, "eval_logits/chosen": 1.162959337234497, "eval_logits/rejected": 1.6319952011108398, "eval_logps/chosen": -537.3916015625, "eval_logps/ref_chosen": -79.05104064941406, "eval_logps/ref_rejected": -86.79793548583984, "eval_logps/rejected": -675.0609741210938, "eval_loss": 0.621041476726532, "eval_rewards/accuracies": 0.712756872177124, "eval_rewards/chosen": -0.6303684711456299, "eval_rewards/margins": 0.17634743452072144, "eval_rewards/rejected": -0.8067159056663513, "eval_runtime": 43.1088, "eval_samples": 2339, "eval_samples_per_second": 54.258, "eval_steps_per_second": 1.717 }