Retrieval-augmented generation (RAG)

A technique that retrieves relevant documents at question time and gives them to a language model, so its answer is grounded in those sources.

A RAG system splits documents into chunks, indexes them (often as embeddings), finds the passages most relevant to each question and includes them in the prompt with instructions to cite them. The term comes from a 2020 Facebook AI Research paper.

RAG keeps answers current and checkable without retraining the model. See how RAG works.

#Glossary #Llms #Developers