quiz · 4 min

Retrieval & RAG Check

  1. What is the core reason RAG reduces hallucinations compared to pure parametric recall?
  2. In the tiny retriever built in Lesson 15, how does the score() function select the best document?
  3. According to the RAG failure modes lesson, what is the metric for measuring whether relevant chunks appear in the top-k results?
  4. What does the RAG failure modes lesson say a well-designed RAG system should do when the retriever finds no relevant source?
  5. The 'lost in the middle' phenomenon described in Lesson 16 means that pasting in many retrieved chunks: