arXiv · Computation and Language· Jongwook Yoon, Jongwon Lim, Sungjib Lim, Woojin Cho, Yohan Jo·· 4 hr agoAI score43
LLM 如何改变否定条件下的预测？
How Do LLMs Change Predictions Under Negation?
AI brief
研究在自建的否定评测基准上测试近期开源与闭源 LLM，发现 37-71% 的情况下模型在否定问句下仍重复原答案（如"What is not the capital of Spain?
Source: arXiv · Computation and Language · arxiv.org