passdrill
AI & LLM Engineering · Evaluation, Testing & Safety · Card 008/010 hard

An evaluation framework for a retrieval-augmented generation (RAG) system reports three separate scores for a generated answer: a score for whether the retrieved passages are relevant to the question, a score for whether the answer directly addresses the question asked, and a score for whether each claim made in the generated answer is actually supported by the retrieved passages. Which of these is the 'faithfulness' (or 'groundedness') score, and what does it specifically measure?

  1. The relevance-of-retrieved-passages score is faithfulness, because a faithful answer is defined entirely by how relevant the retrieval step's passages were to the original question, regardless of what the generated answer actually claims
  2. The answer-addresses-the-question score is faithfulness, because a faithful answer is defined entirely by how directly it addresses the literal wording of the question, regardless of whether its claims are grounded in the retrieved passages
  3. Faithfulness cannot be measured separately from the other two scores, since retrieval relevance, answer relevance, and groundedness in retrieved passages are mathematically the same quantity computed three different ways
  4. The claims-supported-by-retrieved-passages score is faithfulness; it specifically measures whether the claims made in the generated answer can be traced back to and supported by the retrieved context, independent of whether that context was relevant to the question or whether the answer addresses the question asked
Next card → Shuffle