Towards Fully Exploiting LLM Internal States to Enhance Knowledge Boundary Perception

Open in new window