Questioning the Explainability of Large Language Models

Large language models' explainability is sensitive to training randomness, challenging simple explanations.