Talking to Yourself: Defying Forgetting in Large Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914344618426368 |
|---|---|
| author | Sun, Yutao Chen, Mingshuai Zhao, Tiancheng Miao, Phillip Zhang, Zilun Shen, Haozhan Zhu, Ruizhe Yin, Jianwei |
| author_facet | Sun, Yutao Chen, Mingshuai Zhao, Tiancheng Miao, Phillip Zhang, Zilun Shen, Haozhan Zhu, Ruizhe Yin, Jianwei |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_20162 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Talking to Yourself: Defying Forgetting in Large Language Models Sun, Yutao Chen, Mingshuai Zhao, Tiancheng Miao, Phillip Zhang, Zilun Shen, Haozhan Zhu, Ruizhe Yin, Jianwei Computation and Language Artificial Intelligence 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. |
| title | Talking to Yourself: Defying Forgetting in Large Language Models |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2602.20162 |