RAG

Definitions

1. Retrieval-Augmented Generation — a technique that enhances LLM responses by first retrieving relevant documents from an external knowledge base and including them in the prompt context. Reduces hallucinations by grounding the model in actual data rather than relying on memorized training data.

   In plain English: A technique where an AI looks up real information in a database before answering your question, rather than just guessing from memory — like a student who's allowed to check their notes during the exam.

Etymology

2020

Facebook AI Research publishes 'Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks,' coining the term

2023

RAG becomes the default pattern for building LLM applications that need current or domain-specific knowledge

2024

The RAG ecosystem explodes — vector databases, chunking strategies, and retrieval frameworks become a major software category

Related Terms

• Hallucination
• Prompt Engineering
• Embedding
• Context Window
• grounding
• retrieval
• Dual Encoder
• Knowledge Graph
• Retrieval Augmented Generation