AI RESEARCH
Break the Brake, Not the Wheel: Untargeted Jailbreak via Entropy Maximization
arXiv CS.AI
•
ArXi:2605.10764v1 Announce Type: cross Recent studies show that gradient-based universal image jailbreaks on vision-language models (VLMs) exhibit little or no cross-model transferability, casting doubt on the feasibility of transferable multimodal jailbreaks. We revisit this