A Self-Evolving Multi-Agent Framework Defense against LLM Jailbreak Attacks

arXiv:2608.26008v1 Announce Type: cross Abstract: Large language models (LLMs) remain vulnerable to jailbreak attacks that exploit techniques such as role-playing, obfuscation, code transformation, and multi-step indirection to elicit harmful outputs. As jailbreak strategies keep emerging, defenses have proliferated in an ongoing cat-and-mouse game, yet most remain static: their safety behavior is fixed at deployment, so they cannot accumulate defensive experience or adapt to unseen strategies. We propose a self-evolving test-time defense built around a persistent, cross-interaction rule memory: when an attack succeeds, the framework abstracts that failure into a method-level rule capturing the structural attack wrapper rather than the harmful topic, and reuses it against future inputs. Because rules are method-level, one induced rule generalizes across an entire attack family, and the label space expands as novel wrappers appear. The mechanism operates entirely through external memory and prompting, with no parameter updates, and applies to both open-weight and black-box API models. We realize it as four cooperating modules, but the contribution is the memory-based adaptation mechanism, not the module decomposition. Across four black-box jailbreak families and multiple models, our method substantially reduces attack success rates while preserving benign utility, remains robust under an adaptive composite-wrapper attack, and does not increase over-refusal as the memory grows.

Abstract:Large language models (LLMs) remain vulnerable to jailbreak attacks that exploit techniques such as role-playing, obfuscation, code transformation, and multi-step indirection to elicit harmful outputs. As jailbreak strategies keep emerging, defenses have proliferated in an ongoing cat-and-mouse game, yet most remain static: their safety behavior is fixed at deployment, so they cannot accumulate defensive experience or adapt to unseen strategies. We propose a self-evolving test-time defense built around a persistent, cross-interaction rule memory: when an attack succeeds, the framework abstracts that failure into a method-level rule capturing the structural attack wrapper rather than the harmful topic, and reuses it against future inputs. Because rules are method-level, one induced rule generalizes across an entire attack family, and the label space expands as novel wrappers appear. The mechanism operates entirely through external memory and prompting, with no parameter updates, and applies to both open-weight and black-box API models. We realize it as four cooperating modules, but the contribution is the memory-based adaptation mechanism, not the module decomposition. Across four black-box jailbreak families and multiple models, our method substantially reduces attack success rates while preserving benign utility, remains robust under an adaptive composite-wrapper attack, and does not increase over-refusal as the memory grows.

Cryptography and Security (cs.CR); Computation and Language (cs.CL)

arXiv:2608.26008 [cs.CR]

(or arXiv:2608.26008v1 [cs.CR] for this version)

https://doi.org/10.48550/arXiv.2608.26008 Focus to learn more arXiv-issued DOI via DataCite (pending registration)

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

analiza aiqbench

A Self-Evolving Multi-Agent Framework Defense against LLM Jailbreak Attacks

Subiecte conexe