一作ACL 2026Findings· Findings of ACL
Can LLM s Hear the Dogwhistle?
Yifan Liu · Yi Lin · Xinwei Guo · Ziwei Wang · Jiaxin Zhang · Guanhua Chen · Haiyan Wu · an Jiaotong University) · Xin Yao · Xuetao Wei
摘要
With the widespread deployment of large language models (LLMs), existing safety benchmarks remain largely focused on explicitly harmful content, overlooking context-dependent expressions such as dogwhistles, the language that conveys harmful intent while appearing benign on the surface. To addres...