Explaining LLMs for RAG and Summarization | Towards Data Science

A fast and low-resource method using similarity-based attribution