"Vocabulary: Retrieval and RAG"

1. What is the main purpose of an embedding model in the retrieval process?

To directly answer user queries To convert text into fixed-length vectors To store documents and metadata To chunk documents into passages

2. What does the term 'top-k' refer to in the context of retrieval?

The number of documents in the corpus The maximum length of a generated answer The number of nearest chunks returned for a prompt The threshold for similarity scores

3. What is the consequence of 'over-matching' during the retrieval process?

It results in the correct answer being found It allows retrieval to return more relevant chunks It can lead to using the wrong source for generating answers It prevents the model from using any sources at all

4. Why is citation important in the context of RAG?

It helps in generating longer answers It provides a reference for accountability It is used for re-indexing documents It improves the efficiency of embedding

Continue: "Vocabulary: Agents, Tools, and MCP" → Review this lesson Retake quiz