8 papers
Break the Brake, Not the Wheel: Untargeted Jailbreak via Entropy Maximization
Mengqi He, Xinyu Tian, Xin Shen +6
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 feasibili…
High-Entropy Tokens as Multimodal Failure Points in Vision-Language Models
Mengqi He, Xinyu Tian, Xin Shen +4
Vision-language models (VLMs) achieve remarkable performance but remain vulnerable to adversarial attacks. Entropy, as a measure of model uncertainty, is highly correlated with VLM…
All Roads Lead to Rome: Incentivizing Divergent Thinking in Vision-Language Models
Xinyu Tian, Shu Zou, Zhaoyuan Yang +3
Recent studies have demonstrated that Reinforcement Learning (RL), notably Group Relative Policy Optimization (GRPO), can intrinsically elicit and enhance the reasoning capabilitie…
More Thought, Less Accuracy? On the Dual Nature of Reasoning in Vision-Language Models
Xinyu Tian, Shu Zou, Zhaoyuan Yang +5
Reasoning has emerged as a pivotal capability in Large Language Models (LLMs). Through Reinforcement Learning (RL), typically Group Relative Policy Optimization (GRPO), these model…
Unlocking Vision-Language Models for Video Anomaly Detection via Fine-Grained Prompting
Shu Zou, Xinyu Tian, Lukas Wesemann +3
Prompting has emerged as a practical way to adapt frozen vision-language models (VLMs) for video anomaly detection (VAD). Yet, existing prompts are often overly abstract, overlooki…
Identifying and Mitigating Position Bias of Multi-image Vision-Language Models
Xinyu Tian, Shu Zou, Zhaoyuan Yang +1
The evolution of Large Vision-Language Models (LVLMs) has progressed from single to multi-image reasoning. Despite this advancement, our findings indicate that LVLMs struggle to ro…