SEAS: Self-Evolving Adversarial Safety Optimization for Large Language Models
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_ | 1866909437339369472 |
|---|---|
| author | Diao, Muxi Li, Rumei Liu, Shiyang Liao, Guogang Wang, Jingang Cai, Xunliang Xu, Weiran |
| author_facet | Diao, Muxi Li, Rumei Liu, Shiyang Liao, Guogang Wang, Jingang Cai, Xunliang Xu, Weiran |
| contents | As large language models (LLMs) continue to advance in capability and influence, ensuring their security and preventing harmful outputs has become crucial. A promising approach to address these concerns involves training models to automatically generate adversarial prompts for red teaming. However, the evolving subtlety of vulnerabilities in LLMs challenges the effectiveness of current adversarial methods, which struggle to specifically target and explore the weaknesses of these models. To tackle these challenges, we introduce the $\mathbf{S}\text{elf-}\mathbf{E}\text{volving }\mathbf{A}\text{dversarial }\mathbf{S}\text{afety }\mathbf{(SEAS)}$ optimization framework, which enhances security by leveraging data generated by the model itself. SEAS operates through three iterative stages: Initialization, Attack, and Adversarial Optimization, refining both the Red Team and Target models to improve robustness and safety. This framework reduces reliance on manual testing and significantly enhances the security capabilities of LLMs. Our contributions include a novel adversarial framework, a comprehensive safety dataset, and after three iterations, the Target model achieves a security level comparable to GPT-4, while the Red Team model shows a marked increase in attack success rate (ASR) against advanced models. Our code and datasets are released at https://SEAS-LLM.github.io/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_02632 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | SEAS: Self-Evolving Adversarial Safety Optimization for Large Language Models Diao, Muxi Li, Rumei Liu, Shiyang Liao, Guogang Wang, Jingang Cai, Xunliang Xu, Weiran Computation and Language Artificial Intelligence As large language models (LLMs) continue to advance in capability and influence, ensuring their security and preventing harmful outputs has become crucial. A promising approach to address these concerns involves training models to automatically generate adversarial prompts for red teaming. However, the evolving subtlety of vulnerabilities in LLMs challenges the effectiveness of current adversarial methods, which struggle to specifically target and explore the weaknesses of these models. To tackle these challenges, we introduce the $\mathbf{S}\text{elf-}\mathbf{E}\text{volving }\mathbf{A}\text{dversarial }\mathbf{S}\text{afety }\mathbf{(SEAS)}$ optimization framework, which enhances security by leveraging data generated by the model itself. SEAS operates through three iterative stages: Initialization, Attack, and Adversarial Optimization, refining both the Red Team and Target models to improve robustness and safety. This framework reduces reliance on manual testing and significantly enhances the security capabilities of LLMs. Our contributions include a novel adversarial framework, a comprehensive safety dataset, and after three iterations, the Target model achieves a security level comparable to GPT-4, while the Red Team model shows a marked increase in attack success rate (ASR) against advanced models. Our code and datasets are released at https://SEAS-LLM.github.io/. |
| title | SEAS: Self-Evolving Adversarial Safety Optimization for Large Language Models |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2408.02632 |