让大模型自己和自己对话，防止微调时遗忘通用知识。
Talking to Yourself: Defying Forgetting in Large Language Models
- 用模型自动生成对话数据，混合进任务数据微调。
- 50个场景中40个表现优于基线，保持原始模型水平。
- 适合需要稳定性能的工业级模型微调场景。
在窄域任务数据上微调大语言模型时，灾难性遗忘仍是主要挑战，常导致通用知识与推理能力下降。本文提出SA-SFT，一种轻量级自增强方法：在微调前让模型生成自对话，将生成数据与任务数据混合，无需修改优化或训练计划。尽管不依赖外部数据或额外调参，该方法始终缓解遗忘问题，并提升领域内性能。在50个评估场景中，其表现与原模型相当，在40个场景中优于常见基线（如层冻结、外部数据混合）。基于实证结果，我们进一步提出理论分析，指出遗忘部分源于风格引发的参数漂移，而自生成数据带来的自对齐可有效抑制该效应。结果表明，自增强是一种无需灾难性遗忘即可实现鲁棒模型适配的简单高效机制。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。