Answer Relevance
Answer relevance measures whether a response actually addresses the question that was asked, rather than drifting into related-but-off material. It catches the failure where an answer is true and well-sourced but doesn't answer what the user wanted.
RAG & RetrievalAI Evaluation & Reliability
An answer can be grounded in good sources and still miss the point — responding to a nearby question, burying the answer in caveats, or summarizing the documents instead of answering. Answer relevance scores that directly: does the response address what was actually asked?
It’s the third leg of RAG evaluation, and the one closest to user experience. Context relevance checks the retrieval, faithfulness checks that the model stuck to the sources, and answer relevance checks that the result is on-target for the question. A system needs all three: miss any one and the answer is wrong in a different way — irrelevant context, hallucinated content, or a fluent reply to the wrong question.