作者位次 #2/6ACL 2026Poster· Natural Language Processing
Beyond Detection: Evaluating Fallacy Awareness of LLM s in Interactive Scenarios
Conghui Niu · Ningxin Wu · Ziran Zhao · Dong Yu · Chen Kang · Pengyuan Liu
摘要
Large Language Models (LLMs) often fail to recognize fallacious reasoning in real-world interactions, despite strong performance on static fallacy detection tasks. We define this ability as fallacy awareness , the capacity to autonomously perceive and resist fallacies in dynamic, pragmatic contex...