1
StructBreak: Structural Cognitive Overload-Induced Safety Failures in MLLMs
多模态大模型因结构认知过载引发安全漏洞，ACL 2026最新研究发现。
arXiv:2605.25534v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) excel at structural reasoning yet suffer from a sharp logical…