{ "epoch": 1.0, "eval_beta_dpo/beta": 0.0010000000474974513, "eval_beta_dpo/beta_margin_grad_mean": -0.44149303436279297, "eval_beta_dpo/beta_margin_grad_std": 0.09785857796669006, "eval_beta_dpo/beta_margin_mean": 0.24613134562969208, "eval_beta_dpo/beta_margin_std": 0.41966694593429565, "eval_beta_dpo/beta_used": 0.0010000000474974513, "eval_beta_dpo/beta_used_raw": -10.40408706665039, "eval_beta_dpo/gap_mean": 421.1994323730469, "eval_beta_dpo/gap_std": 381.64532470703125, "eval_beta_dpo/loss_margin_mean": 246.13131713867188, "eval_beta_dpo/mask_keep_frac": 1.0, "eval_logits/chosen": -2.795379400253296, "eval_logits/rejected": -2.8268661499023438, "eval_loss": 0.6000753045082092, "eval_runtime": 37.6049, "eval_samples": 2339, "eval_samples_per_second": 62.199, "eval_steps_per_second": 1.968 }