Retrieval Augmented Generation
In plain terms
Grounding LLM answers by retrieving relevant documents first.
Going deeper
RAG fetches relevant information from a knowledge source and feeds it to the model as context, producing grounded, current answers without retraining. It combines search with generation.
Under the hood
Retrieval quality (chunking, embeddings, reranking) drives accuracy; reduces hallucination but doesn't eliminate it.
Examples
Chatbots over docs, semantic search + LLM, citations, enterprise Q&A.
Think of it like...
Like an open-book exam where the model looks up facts before answering.