SOURCE-LINKED INTELLIGENCE

DIVA: Exploiting Cross-Step Conditional Propagation for Visual Jailbreaks in Discrete Diffusion Vision-Language Models

arXiv · AI, language, vision and robotics · article · Sep 1, 2026 · UTC

Large vision-language models (VLMs) are increasingly deployed in safety-critical settings, yet existing visual jailbreak research has focused almost exclusively on autoregressive architectures, leaving an important emerging family unstudied: multimodal discrete diffusion vision-language models (dVLMs). We identify a vulnerability specific to diffusion generation: because the visual embedding conditions every reverse denoising step rather than acting as a one-time prefix, adversarial visual semantics are repeatedly propagated and amplified across the generation trajectory, a phenomenon we term

Read original source ↗ Open in workspace

recordType

paper

region

Global