一作ACL 2026Poster· Natural Language Processing
False F riends in the Shell: Unveiling the Emoticon Semantic Confusion in Large Language Models
Weipeng Jiang · Xiaoyu Zhang · Juan Zhai · Shiqing Ma · Chao Shen · Yang Liu
摘要
Emoticons are widely used in digital communication to convey affective intent, yet their safety implications for Large Language Models (LLMs) remain largely unexplored. In this paper, we identify emoticon semantic confusion, a vulnerability where LLMs misinterpret ASCII-based emoticons to perform...