作者位次 #3/6ACL 2026Findings· Findings of ACL
VLMG uard- R 1: Proactive Safety Alignment for VLM s via Reasoning-Driven Prompt Optimization
Menglan Chen · Xianghe Pang · Jingjing Dong · WenHao Wang · Yaxin Du · Siheng Chen
摘要
Aligning Vision-Language Models (VLMs) with safety standards is essential to mitigate risks arising from their multimodal complexity, where integrating vision and language unveils subtle threats beyond the reach of conventional safeguards. Inspired by the insight that reasoning across modalities...