Safeguard Fine-Tuned LLMs Through Pre- and Post-Tuning Model Merging
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916922455490560 |
|---|---|
| author | Farn, Hua Su, Hsuan Kumar, Shachi H Sahay, Saurav Chen, Shang-Tse Lee, Hung-yi |
| author_facet | Farn, Hua Su, Hsuan Kumar, Shachi H Sahay, Saurav Chen, Shang-Tse Lee, Hung-yi |
| contents | Fine-tuning large language models (LLMs) for downstream tasks often leads to catastrophic forgetting, notably degrading the safety of originally aligned models. While some existing methods attempt to restore safety by incorporating additional safety data, the quality of such data typically falls short of that used in the original alignment process. Moreover, these high-quality safety datasets are generally inaccessible, making it difficult to fully recover the model's original safety. We ask: How can we preserve safety while improving downstream task performance without additional safety data? We show that simply merging the weights of pre- and post-fine-tuned models effectively mitigates safety degradation while enhancing performance. Experiments across different downstream tasks and models validate the method's practicality and effectiveness. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_19512 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Safeguard Fine-Tuned LLMs Through Pre- and Post-Tuning Model Merging Farn, Hua Su, Hsuan Kumar, Shachi H Sahay, Saurav Chen, Shang-Tse Lee, Hung-yi Computation and Language Fine-tuning large language models (LLMs) for downstream tasks often leads to catastrophic forgetting, notably degrading the safety of originally aligned models. While some existing methods attempt to restore safety by incorporating additional safety data, the quality of such data typically falls short of that used in the original alignment process. Moreover, these high-quality safety datasets are generally inaccessible, making it difficult to fully recover the model's original safety. We ask: How can we preserve safety while improving downstream task performance without additional safety data? We show that simply merging the weights of pre- and post-fine-tuned models effectively mitigates safety degradation while enhancing performance. Experiments across different downstream tasks and models validate the method's practicality and effectiveness. |
| title | Safeguard Fine-Tuned LLMs Through Pre- and Post-Tuning Model Merging |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2412.19512 |