Retrieval-augmented generation

RAG combines LLMs with external knowledge bases. The model retrieves relevant documents, then uses them to generate grounded, factual responses.

• RAG grounds LLM responses in external data.
• It reduces hallucinations by providing factual context.