AI RESEARCH
Do LLMs Really Know What They Don't Know? Internal States Mainly Reflect Knowledge Recall Rather Than Truthfulness
arXiv CS.CL
•
ArXi:2510.09033v2 Announce Type: replace Recent work suggests that LLMs "know what they don't know", positing that hallucinated and factually correct outputs arise from distinct internal processes and can. therefore. be distinguished using internal signals. However, hallucinations have multifaceted causes: beyond simple knowledge gaps, they can emerge from