JailbreakLens: Visual Analysis of Jailbreak Attacks Against 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_ | 1866912417631436800 |
|---|---|
| author | Feng, Yingchaojie Chen, Zhizhang Kang, Zhining Wang, Sijia Tian, Haoyu Zhang, Wei Zhu, Minfeng Chen, Wei |
| author_facet | Feng, Yingchaojie Chen, Zhizhang Kang, Zhining Wang, Sijia Tian, Haoyu Zhang, Wei Zhu, Minfeng Chen, Wei |
| contents | The proliferation of large language models (LLMs) has underscored concerns regarding their security vulnerabilities, notably against jailbreak attacks, where adversaries design jailbreak prompts to circumvent safety mechanisms for potential misuse. Addressing these concerns necessitates a comprehensive analysis of jailbreak prompts to evaluate LLMs' defensive capabilities and identify potential weaknesses. However, the complexity of evaluating jailbreak performance and understanding prompt characteristics makes this analysis laborious. We collaborate with domain experts to characterize problems and propose an LLM-assisted framework to streamline the analysis process. It provides automatic jailbreak assessment to facilitate performance evaluation and support analysis of components and keywords in prompts. Based on the framework, we design JailbreakLens, a visual analysis system that enables users to explore the jailbreak performance against the target model, conduct multi-level analysis of prompt characteristics, and refine prompt instances to verify findings. Through a case study, technical evaluations, and expert interviews, we demonstrate our system's effectiveness in helping users evaluate model security and identify model weaknesses. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_08793 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | JailbreakLens: Visual Analysis of Jailbreak Attacks Against Large Language Models Feng, Yingchaojie Chen, Zhizhang Kang, Zhining Wang, Sijia Tian, Haoyu Zhang, Wei Zhu, Minfeng Chen, Wei Cryptography and Security Computation and Language Human-Computer Interaction The proliferation of large language models (LLMs) has underscored concerns regarding their security vulnerabilities, notably against jailbreak attacks, where adversaries design jailbreak prompts to circumvent safety mechanisms for potential misuse. Addressing these concerns necessitates a comprehensive analysis of jailbreak prompts to evaluate LLMs' defensive capabilities and identify potential weaknesses. However, the complexity of evaluating jailbreak performance and understanding prompt characteristics makes this analysis laborious. We collaborate with domain experts to characterize problems and propose an LLM-assisted framework to streamline the analysis process. It provides automatic jailbreak assessment to facilitate performance evaluation and support analysis of components and keywords in prompts. Based on the framework, we design JailbreakLens, a visual analysis system that enables users to explore the jailbreak performance against the target model, conduct multi-level analysis of prompt characteristics, and refine prompt instances to verify findings. Through a case study, technical evaluations, and expert interviews, we demonstrate our system's effectiveness in helping users evaluate model security and identify model weaknesses. |
| title | JailbreakLens: Visual Analysis of Jailbreak Attacks Against Large Language Models |
| topic | Cryptography and Security Computation and Language Human-Computer Interaction |
| url | https://arxiv.org/abs/2404.08793 |