Boosting Jailbreak Attack with Momentum
Fuente:
arXiv
Guardado en:
| Autores principales: | , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866912253953966080 |
|---|---|
| author | Zhang, Yihao Wei, Zeming |
| author_facet | Zhang, Yihao Wei, Zeming |
| contents | Large Language Models (LLMs) have achieved remarkable success across diverse tasks, yet they remain vulnerable to adversarial attacks, notably the well-known jailbreak attack. In particular, the Greedy Coordinate Gradient (GCG) attack has demonstrated efficacy in exploiting this vulnerability by optimizing adversarial prompts through a combination of gradient heuristics and greedy search. However, the efficiency of this attack has become a bottleneck in the attacking process. To mitigate this limitation, in this paper we rethink the generation of the adversarial prompts through an optimization lens, aiming to stabilize the optimization process and harness more heuristic insights from previous optimization iterations. Specifically, we propose the \textbf{M}omentum \textbf{A}ccelerated G\textbf{C}G (\textbf{MAC}) attack, which integrates a momentum term into the gradient heuristic to boost and stabilize the random search for tokens in adversarial prompts. Experimental results showcase the notable enhancement achieved by MAC over baselines in terms of attack success rate and optimization efficiency. Moreover, we demonstrate that MAC can still exhibit superior performance for transfer attacks and models under defense mechanisms. Our code is available at https://github.com/weizeming/momentum-attack-llm. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_01229 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Boosting Jailbreak Attack with Momentum Zhang, Yihao Wei, Zeming Machine Learning Artificial Intelligence Computation and Language Cryptography and Security Optimization and Control Large Language Models (LLMs) have achieved remarkable success across diverse tasks, yet they remain vulnerable to adversarial attacks, notably the well-known jailbreak attack. In particular, the Greedy Coordinate Gradient (GCG) attack has demonstrated efficacy in exploiting this vulnerability by optimizing adversarial prompts through a combination of gradient heuristics and greedy search. However, the efficiency of this attack has become a bottleneck in the attacking process. To mitigate this limitation, in this paper we rethink the generation of the adversarial prompts through an optimization lens, aiming to stabilize the optimization process and harness more heuristic insights from previous optimization iterations. Specifically, we propose the \textbf{M}omentum \textbf{A}ccelerated G\textbf{C}G (\textbf{MAC}) attack, which integrates a momentum term into the gradient heuristic to boost and stabilize the random search for tokens in adversarial prompts. Experimental results showcase the notable enhancement achieved by MAC over baselines in terms of attack success rate and optimization efficiency. Moreover, we demonstrate that MAC can still exhibit superior performance for transfer attacks and models under defense mechanisms. Our code is available at https://github.com/weizeming/momentum-attack-llm. |
| title | Boosting Jailbreak Attack with Momentum |
| topic | Machine Learning Artificial Intelligence Computation and Language Cryptography and Security Optimization and Control |
| url | https://arxiv.org/abs/2405.01229 |