Fetching the paper…

InternalInspector $I^2$: Robust Confidence Estimation in LLMs through Internal States · Around