Rethinking Jailbreak Detection of Large Vision Language Models with Representational Contrastive Scoring

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Hua, Peichun, Li, Hao, Shi, Shanghao, Yu, Zhiyuan, Zhang, Ning
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914487699767296
author Hua, Peichun
Li, Hao
Shi, Shanghao
Yu, Zhiyuan
Zhang, Ning
author_facet Hua, Peichun
Li, Hao
Shi, Shanghao
Yu, Zhiyuan
Zhang, Ning
contents Large Vision-Language Models (LVLMs) are vulnerable to a growing array of multimodal jailbreak attacks, necessitating defenses that are both generalizable to novel threats and efficient for practical deployment. Many current strategies fall short, either targeting specific attack patterns, which limits generalization, or imposing high computational overhead. While lightweight anomaly-detection methods offer a promising direction, we find that their common one-class design tends to confuse unseen benign inputs with malicious ones, leading to unreliable over-rejection. To address this, we propose Representational Contrastive Scoring (RCS), a framework built on a key insight: the most potent safety signals reside within the LVLM's own internal representations. Our approach inspects the internal geometry of these representations, learning a lightweight projection to maximally separate benign and malicious inputs in safety-critical layers. This enables a simple yet powerful contrastive score that differentiates true malicious intent from mere distribution shift. Our instantiations, MCD (Mahalanobis Contrastive Detection) and KCD (K-nearest Contrastive Detection), achieve state-of-the-art performance on a challenging evaluation protocol designed to test generalization to unseen attack types. This work demonstrates that effective jailbreak detection can be achieved by applying simple, interpretable statistical methods to the internal representations, offering a practical path towards safer LVLM deployment.
format Preprint
id arxiv_https___arxiv_org_abs_2512_12069
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Rethinking Jailbreak Detection of Large Vision Language Models with Representational Contrastive Scoring
Hua, Peichun
Li, Hao
Shi, Shanghao
Yu, Zhiyuan
Zhang, Ning
Cryptography and Security
Artificial Intelligence
Computation and Language
Machine Learning
Large Vision-Language Models (LVLMs) are vulnerable to a growing array of multimodal jailbreak attacks, necessitating defenses that are both generalizable to novel threats and efficient for practical deployment. Many current strategies fall short, either targeting specific attack patterns, which limits generalization, or imposing high computational overhead. While lightweight anomaly-detection methods offer a promising direction, we find that their common one-class design tends to confuse unseen benign inputs with malicious ones, leading to unreliable over-rejection. To address this, we propose Representational Contrastive Scoring (RCS), a framework built on a key insight: the most potent safety signals reside within the LVLM's own internal representations. Our approach inspects the internal geometry of these representations, learning a lightweight projection to maximally separate benign and malicious inputs in safety-critical layers. This enables a simple yet powerful contrastive score that differentiates true malicious intent from mere distribution shift. Our instantiations, MCD (Mahalanobis Contrastive Detection) and KCD (K-nearest Contrastive Detection), achieve state-of-the-art performance on a challenging evaluation protocol designed to test generalization to unseen attack types. This work demonstrates that effective jailbreak detection can be achieved by applying simple, interpretable statistical methods to the internal representations, offering a practical path towards safer LVLM deployment.
title Rethinking Jailbreak Detection of Large Vision Language Models with Representational Contrastive Scoring
topic Cryptography and Security
Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2512.12069