Explain RAG step 5:
In this step, the system takes two distinct inputs and merges them into a single coherent structure:
• The Original Query: The initial question or prompt provided in Step 1.
• The Retrieved Context: The specific "Similar docs" or text chunks found during the Similarity Search in Step 4. These are the external facts that the LLM was not originally trained on (such as private documents or real-time news).
Why this step matters
The goal of Step 5 is to construct a Rich Prompt. Instead of just asking the LLM a question blindly, the system essentially says:
"Using the following reference material: [Retrieved Context], please answer this question: [Query]."
