{ "epoch": 1.0, "eval_beta_dpo/beta": 0.011150650680065155, "eval_beta_dpo/beta_margin_grad_mean": -0.47102445363998413, "eval_beta_dpo/beta_margin_grad_std": 0.049201950430870056, "eval_beta_dpo/beta_margin_mean": 1.640921711921692, "eval_beta_dpo/beta_margin_std": 2.0729942321777344, "eval_beta_dpo/beta_used": 0.011150650680065155, "eval_beta_dpo/beta_used_raw": -3.504255771636963, "eval_beta_dpo/gap_mean": 147.1534881591797, "eval_beta_dpo/gap_std": 168.50018310546875, "eval_beta_dpo/loss_margin_mean": 87.08258056640625, "eval_beta_dpo/mask_keep_frac": 1.0, "eval_logits/chosen": -0.2772989273071289, "eval_logits/rejected": -0.2549748718738556, "eval_loss": 0.7894024848937988, "eval_runtime": 40.1139, "eval_samples": 2339, "eval_samples_per_second": 58.309, "eval_steps_per_second": 1.845 }