Retrieval-Augmented Generation (RAG)
An architectural technique that enriches the prompts of large language models (LLMs) with external facts retrieved from vector databases or files in real-time, ensuring highly accurate, contextual, and hallucination-free responses.