So, What’s Actually Happening in RAG?

In the previous exercise, you sequenced the steps for a RAG pipeline that finds the right information from home appliance manuals.

Assume you assembled the RAG pipeline as follows:

• Load manual PDFs
• Split with chunk_size=1000, chunk_overlap=200
• Create embeddings and index
• Retrieve using similarity_search(query, k=3)
• Join retrieved chunks into a context string to send to an LLM

Which of the following options correctly describe what the pipeline does?