What is the primary goal of Contextual Retrieval in a RAG pipeline?	To improve retrieval accuracy by adding context to individual text chunks.
What pre-processing step defines Contextual Retrieval?	Sending a text chunk and the source document to an LLM (like Claude) to generate contextual information.
What is the final product created after the LLM generates context for a chunk?	A "contextualized chunk," which combines the original chunk text and the added context.
What problem does Contextual Retrieval aim to fix regarding document chunking?	That individual chunks often lack the necessary context from the original source document.
If the source document is too large to fit into a single prompt for contextualization, what is a viable strategy?	Include strategic, smaller chunks instead of the entire document.
When contextualizing a specific chunk in a large document, what types of chunks should be included?	Starter chunks (for overall document summary) and chunks immediately preceding the target chunk.
What specific task is the LLM asked to perform when contextualizing a chunk?	To write text that situates or places the chunk within the context of the larger document.
What is the final input used for the vector index and BM25 index in this technique?	The contextualized chunk.