与自己对话：在大型语言模型中克服遗忘
Talking to Yourself: Defying Forgetting in Large Language Models
AI总结：
本文提出SA-SFT方法，通过让LLM生成自我对话来缓解微调过程中的灾难性遗忘，提升领域性能并优于现有基线方法。
摘要：
在对大型语言模型（LLMs）进行微调时，灾难性遗忘仍然是一个主要挑战，通常会降低其通用知识和推理能力。我们提出了SA-SFT，一种轻量级的自我增强流程，在微调前让LLM生成自我对话，生成的自我创作数据与任务数据混合，而无需修改优化或训练计划。尽管不需要外部数据或额外调优，SA-SFT一致地缓解了灾难性遗忘并提高了领域内的性能。在50个评估场景中，它保持的性能与原始模型相当，并在40种情况下取得最佳结果，优于常见的基线方法，如层冻结和外部数据混合。基于这些经验发现，我们进一步提出了一个理论分析，表明遗忘可能部分源于风格引起的参数漂移，而通过自我生成的数据进行自我对齐是一种有效对抗这种效应的方法。总体而言，我们的结果表明，自我增强为无需产生灾难性遗忘的稳健LLM适应提供了一种简单而有效的方法。
英文摘要：
Catastrophic forgetting remains a major challenge when fine-tuning large language models (LLMs) on narrow, task-specific data, often degrading their general knowledge and reasoning abilities. We propose SA-SFT, a lightweight self-augmentation routine in which an LLM generates self-dialogues prior to fine-tuning, and the resulting self-authored data are mixed with task data without modifying optimization or training schedules. Despite requiring no external data or additional tuning, SA-SFT consistently mitigates catastrophic forgetting while improving in-domain performance. Across 50 evaluation scenarios, it maintains performance comparable to the original model and achieves the best results in 40 cases, outperforming common baselines such as layer freezing and external data mixing. Guided by these empirical findings, we further present a theoretical analysis suggesting that forgetting can partly stem from style-induced parameter drift, and that self-alignment through self-generated data provides an effective means to counteract this effect. Overall, our results indicate that self-augmentation offers a simple and effective mechanism for robust LLM adaptation without incurring catastrophic forgetting.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
