Update the popular RAG notebooks with notes on how to run inference with chat messages. Also provide links to the LLM and RAG pipeline documentation.