JailBound: Jailbreaking Internal Safety Boundaries of Vision-Language Models

TL;DR

JailBound introduces a novel latent space jailbreak framework for vision-language models, effectively exploiting internal safety boundaries to generate targeted attacks and reveal safety vulnerabilities in diverse VLMs.

Contribution

The paper proposes JailBound, a two-stage latent space attack method that improves cross-modal attack effectiveness by approximating safety boundaries and jointly optimizing perturbations.

Findings

Achieves 94.32% success in white-box attacks

Outperforms state-of-the-art methods by 6.17% in success rate

Reveals significant safety risks in current VLMs

Abstract

Vision-Language Models (VLMs) exhibit impressive performance, yet the integration of powerful vision encoders has significantly broadened their attack surface, rendering them increasingly susceptible to jailbreak attacks. However, lacking well-defined attack objectives, existing jailbreak methods often struggle with gradient-based strategies prone to local optima and lacking precise directional guidance, and typically decouple visual and textual modalities, thereby limiting their effectiveness by neglecting crucial cross-modal interactions. Inspired by the Eliciting Latent Knowledge (ELK) framework, we posit that VLMs encode safety-relevant information within their internal fusion-layer representations, revealing an implicit safety decision boundary in the latent space. This motivates exploiting boundary to steer model behavior. Accordingly, we propose JailBound, a novel latent space jailbreak framework comprising two stages: (1) Safety Boundary Probing, which addresses the guidance issue by approximating decision boundary within fusion layer's latent space, thereby identifying optimal perturbation directions towards the target region; and (2) Safety Boundary Crossing, which overcomes the limitations of decoupled approaches by jointly optimizing adversarial perturbations across both image and text inputs. This latter stage employs an innovative mechanism to steer the model's internal state towards policy-violating outputs while maintaining cross-modal semantic consistency. Extensive experiments on six diverse VLMs demonstrate JailBound's efficacy, achieves 94.32% white-box and 67.28% black-box attack success averagely, which are 6.17% and 21.13% higher than SOTA methods, respectively. Our findings expose a overlooked safety risk in VLMs and highlight the urgent need for more robust defenses. Warning: This paper contains potentially sensitive, harmful and offensive content.