Recall which RAG quality metric maps to which failure and which fix.
Retrieval When do you use context relevance? Use it when you need to know whether the retrieved chunks are actually useful for the question. Low context relevance points to query rewrite, filters, chunking, hybrid search, or reranking. Recall vs faithfulness Source absent or source ignored? The distinction saves teams from buying more recall for a generator bug. Triage objection The answer was wrong, so we need a bigger model. Use when the room jumps from bad answer to model upgrade. Your line Maybe, but first we need to know whether the right evidence was retrieved and whether the answer stayed…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in