Retrieval Optimization for RAG Quiz

What you'll be asked

1. What is the primary purpose of implementing a 'chunking' strategy in a RAG system?
2. When using dense vector retrieval, why might a 'Hybrid Search' approach be more effective than vector search alone?
3. What is the main benefit of implementing a Reranking step after the initial retrieval process?
4. If your RAG system is retrieving semantically similar but factually irrelevant documents, which optimization strategy should you prioritize first?
5. In a scenario where you are using a fixed embedding model, how does 'Contextual Retrieval' (prepending document-level summaries to individual chunks) specifically address the 'Lost in the Middle' phenomenon?

Answers and explanations come after you submit.