midLLMOps
What is retrieval augmented generation (RAG)?
Updated May 16, 2026
Short answer
RAG combines LLMs with external knowledge retrieval to improve factual accuracy.
Deep explanation
RAG systems retrieve relevant documents from a knowledge base and inject them into the LLM prompt. This reduces hallucinations and allows models to access up-to-date or domain-specific information without retraining.
Real-world example
Enterprise chatbots using internal company documents for answers.
Common mistakes
- Relying only on model memory without retrieval.
Follow-up questions
- Why is RAG better than fine-tuning?
- What is vector database?