TRAP: Hijacking VLA CoT-Reasoning via Adversarial Patches 文章

ArXiv CS.AI2026-06-03NEWSen作者: Zhengxian Huang, Wenjun Zhu, Haoxuan Qiu, Xiaoyu Ji, Wenyuan Xu
查看原文 →
关系图谱

TRAP: Hijacking VLA CoT-Reasoning via Adversarial Patches · 相关技术

相关技术

TRAPVLA (Vision-Language-Action)chain-of-thought