LLMs know more than they show: On the intrinsic representation of hallucinations October 30, 2024 by Comments