AI and LLMs · Middle · Question 70 of 75
What is a common mistake when building a first RAG prototype?
- 1Never measuring retrieval separately from answer quality
- 2Splitting documents into chunks at all
- 3Using embeddings instead of plain keyword search
Show the answer
Correct answer: Never measuring retrieval separately from answer quality
Explanation
When answers are bad, you need to know whether retrieval missed the passage or generation misused it.