{ "epoch": 1.0, "eval_logits/chosen": -3.01904296875, "eval_logits/rejected": -3.02587890625, "eval_logps/chosen": -382.25, "eval_logps/rejected": -417.0, "eval_loss": 0.5717031359672546, "eval_rewards/accuracies": 0.72119140625, "eval_rewards/chosen": -0.43609619140625, "eval_rewards/margins": 0.5413818359375, "eval_rewards/rejected": -0.9776611328125, "eval_runtime": 38.7296, "eval_samples_per_second": 51.64, "eval_steps_per_second": 0.826 }