RAG (Retrieval-Augmented Generation)

A system architecture that improves LLM outputs by pulling relevant external documents before generating a response.