Hidden You Malicious Goal Into Benign Narratives: Jailbreak Large Language Models through Logic Chain Injection
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_ | 1866916209671274496 |
|---|---|
| author | Wang, Zhilong Cao, Yebo Liu, Peng |
| author_facet | Wang, Zhilong Cao, Yebo Liu, Peng |
| contents | Jailbreak attacks on Language Model Models (LLMs) entail crafting prompts aimed at exploiting the models to generate malicious content. Existing jailbreak attacks can successfully deceive the LLMs, however they cannot deceive the human. This paper proposes a new type of jailbreak attacks which can deceive both the LLMs and human (i.e., security analyst). The key insight of our idea is borrowed from the social psychology - that is human are easily deceived if the lie is hidden in truth. Based on this insight, we proposed the logic-chain injection attacks to inject malicious intention into benign truth. Logic-chain injection attack firstly dissembles its malicious target into a chain of benign narrations, and then distribute narrations into a related benign article, with undoubted facts. In this way, newly generate prompt cannot only deceive the LLMs, but also deceive human. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_04849 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Hidden You Malicious Goal Into Benign Narratives: Jailbreak Large Language Models through Logic Chain Injection Wang, Zhilong Cao, Yebo Liu, Peng Cryptography and Security Artificial Intelligence Jailbreak attacks on Language Model Models (LLMs) entail crafting prompts aimed at exploiting the models to generate malicious content. Existing jailbreak attacks can successfully deceive the LLMs, however they cannot deceive the human. This paper proposes a new type of jailbreak attacks which can deceive both the LLMs and human (i.e., security analyst). The key insight of our idea is borrowed from the social psychology - that is human are easily deceived if the lie is hidden in truth. Based on this insight, we proposed the logic-chain injection attacks to inject malicious intention into benign truth. Logic-chain injection attack firstly dissembles its malicious target into a chain of benign narrations, and then distribute narrations into a related benign article, with undoubted facts. In this way, newly generate prompt cannot only deceive the LLMs, but also deceive human. |
| title | Hidden You Malicious Goal Into Benign Narratives: Jailbreak Large Language Models through Logic Chain Injection |
| topic | Cryptography and Security Artificial Intelligence |
| url | https://arxiv.org/abs/2404.04849 |