What is RAG (Retrieval-Augmented Generation)?
Retrieval-Augmented Generation
A. What is RAG (Retrieval-Augmented Generation)
A method where an AI searches external documents and answers based on their content. It is effective for handling up-to-date information and internal documents, and for reducing hallucinations.
RAG (Retrieval-Augmented Generation) is a technique that searches a database or the web for documents relevant to a question, then includes their content in the prompt so the LLM can answer based on them.
It lets a model handle up-to-date information or proprietary internal data it was never trained on, and because it can cite evidence, it also helps reduce hallucinations. It is a core technology behind AI search and internal AI assistants.