Obliviate: Neutralizing Task-agnostic Backdoors within the Parameter-efficient Fine-tuning Paradigm
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909336900468736 |
|---|---|
| author | Kim, Jaehan Song, Minkyoo Na, Seung Ho Shin, Seungwon |
| author_facet | Kim, Jaehan Song, Minkyoo Na, Seung Ho Shin, Seungwon |
| contents | Parameter-efficient fine-tuning (PEFT) has become a key training strategy for large language models. However, its reliance on fewer trainable parameters poses security risks, such as task-agnostic backdoors. Despite their severe impact on a wide range of tasks, there is no practical defense solution available that effectively counters task-agnostic backdoors within the context of PEFT. In this study, we introduce Obliviate, a PEFT-integrable backdoor defense. We develop two techniques aimed at amplifying benign neurons within PEFT layers and penalizing the influence of trigger tokens. Our evaluations across three major PEFT architectures show that our method can significantly reduce the attack success rate of the state-of-the-art task-agnostic backdoors (83.6%$\downarrow$). Furthermore, our method exhibits robust defense capabilities against both task-specific backdoors and adaptive attacks. Source code will be obtained at https://github.com/obliviateARR/Obliviate. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_14119 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Obliviate: Neutralizing Task-agnostic Backdoors within the Parameter-efficient Fine-tuning Paradigm Kim, Jaehan Song, Minkyoo Na, Seung Ho Shin, Seungwon Computation and Language Artificial Intelligence Cryptography and Security Machine Learning Parameter-efficient fine-tuning (PEFT) has become a key training strategy for large language models. However, its reliance on fewer trainable parameters poses security risks, such as task-agnostic backdoors. Despite their severe impact on a wide range of tasks, there is no practical defense solution available that effectively counters task-agnostic backdoors within the context of PEFT. In this study, we introduce Obliviate, a PEFT-integrable backdoor defense. We develop two techniques aimed at amplifying benign neurons within PEFT layers and penalizing the influence of trigger tokens. Our evaluations across three major PEFT architectures show that our method can significantly reduce the attack success rate of the state-of-the-art task-agnostic backdoors (83.6%$\downarrow$). Furthermore, our method exhibits robust defense capabilities against both task-specific backdoors and adaptive attacks. Source code will be obtained at https://github.com/obliviateARR/Obliviate. |
| title | Obliviate: Neutralizing Task-agnostic Backdoors within the Parameter-efficient Fine-tuning Paradigm |
| topic | Computation and Language Artificial Intelligence Cryptography and Security Machine Learning |
| url | https://arxiv.org/abs/2409.14119 |