What is the first step in the Retrieval-Augmented Generation (RAG) workflow?	Chunking the source text (e.g., a report) by section.
What process is applied to every text chunk created in Step 1?	Generating an embedding for each chunk.
When creating the vector store (Step 3), what two pieces of information must be stored together for each entry?	The vector embedding and the original chunk text (or chunk ID).
Why is it necessary to store the original chunk text alongside its embedding?	To retrieve the actual, meaningful context when the vector database returns a list of similar embeddings.
What is the first action taken when a user submits a question in the RAG workflow?	Generating an embedding for the user's query (the plain string).
How is the relevant information retrieved from the vector store?	By searching the store using the user's query embedding.
What metric is used to measure the relevance of a retrieved chunk to the query?	The distance (e.g., cosine distance).