Why LLMs Say “I Don’t Know” So Badly
中文摘要
大语言模型难以像人类一样承认无知，它们更倾向于产生幻觉而非说“我不知道”。
English Summary
LLMs struggle to admit ignorance, often hallucinating instead of saying "I don't know" like humans.
原文节选
Ask a person a question they know nothing about, and there is a good chance they will simply say: Continue reading on Medium »