Pattern Enhanced Multi-Turn Jailbreaking: Exploiting Structural Vulnerabilities in Large Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914307172728832 |
|---|---|
| author | Nihal, Ragib Amin Wen, Rui Nakadai, Kazuhiro Sakuma, Jun |
| author_facet | Nihal, Ragib Amin Wen, Rui Nakadai, Kazuhiro Sakuma, Jun |
| contents | Large language models (LLMs) remain vulnerable to multi-turn jailbreaking attacks that exploit conversational context to bypass safety constraints gradually. These attacks target different harm categories through distinct conversational approaches. Existing multi-turn methods often rely on heuristic or ad hoc exploration strategies, providing limited insight into underlying model weaknesses. The relationship between conversation patterns and model vulnerabilities across harm categories remains poorly understood. We propose Pattern Enhanced Chain of Attack (PE-CoA), a framework of five conversation patterns to construct multi-turn jailbreaks through natural dialogue. Evaluating PE-CoA on twelve LLMs spanning ten harm categories, we achieve state-of-the-art performance, uncovering pattern-specific vulnerabilities and LLM behavioral characteristics: models exhibit distinct weakness profiles, defense to one pattern does not generalize to others, and model families share similar failure modes. These findings highlight limitations of safety training and indicate the need for pattern-aware defenses. Code available on: https://github.com/Ragib-Amin-Nihal/PE-CoA |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_08859 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Pattern Enhanced Multi-Turn Jailbreaking: Exploiting Structural Vulnerabilities in Large Language Models Nihal, Ragib Amin Wen, Rui Nakadai, Kazuhiro Sakuma, Jun Computation and Language Artificial Intelligence Cryptography and Security Large language models (LLMs) remain vulnerable to multi-turn jailbreaking attacks that exploit conversational context to bypass safety constraints gradually. These attacks target different harm categories through distinct conversational approaches. Existing multi-turn methods often rely on heuristic or ad hoc exploration strategies, providing limited insight into underlying model weaknesses. The relationship between conversation patterns and model vulnerabilities across harm categories remains poorly understood. We propose Pattern Enhanced Chain of Attack (PE-CoA), a framework of five conversation patterns to construct multi-turn jailbreaks through natural dialogue. Evaluating PE-CoA on twelve LLMs spanning ten harm categories, we achieve state-of-the-art performance, uncovering pattern-specific vulnerabilities and LLM behavioral characteristics: models exhibit distinct weakness profiles, defense to one pattern does not generalize to others, and model families share similar failure modes. These findings highlight limitations of safety training and indicate the need for pattern-aware defenses. Code available on: https://github.com/Ragib-Amin-Nihal/PE-CoA |
| title | Pattern Enhanced Multi-Turn Jailbreaking: Exploiting Structural Vulnerabilities in Large Language Models |
| topic | Computation and Language Artificial Intelligence Cryptography and Security |
| url | https://arxiv.org/abs/2510.08859 |