Retrieval-Augmented Generation (RAG) is an artificial intelligence technique designed to enhance the performance of Large Language Models (LLMs). It works by combining the text-generation capabilities of these models with information retrieval from external, authoritative knowledge sources. Essentially, before generating a response, the model consults a specific knowledge base to fetch up-to-date and contextual information.
Despite their power, Large Language Models have limitations. Their knowledge is static, based on the data they were trained on, which can lead to outdated or "stale" information. They are also prone to "hallucinations"—inventing facts that sound plausible but are incorrect. RAG directly addresses these issues by grounding the LLM with real-time, external data.