Retrieval, RAG & Embeddings Retrieval-Augmented Generation (RAG) A technique where an LLM fetches relevant documents before generating an answer. ← Back to the glossary