---
title: "RAG (Retrieval-Augmented Generation)"
slug: "rag"
category: "glossary"
tags: ["glossary", "ai-agents", "llm", "rag", "embeddings"]
status: "stable"
last_updated: 2026-05-14
summary: "RAG augments an LLM's response by retrieving documents from an external store at inference time and injecting them into the prompt as grounding context."
related:
  [
    "[[ai-agents/rag]]",
    "[[glossary/embedding]]",
    "[[glossary/vector-similarity]]",
    "[[glossary/reranker]]",
    "[[glossary/ground-truth]]",
    "[[glossary/hallucination]]",
    "[[glossary/eval-set]]",
  ]
---

## Overview

This page is the atomic definition. Full RAG architecture patterns live at [[ai-agents/rag]].

## Definition

Retrieval-Augmented Generation (RAG) is an inference-time technique that supplements an LLM's parametric knowledge (knowledge baked into weights) with non-parametric knowledge retrieved from an external corpus. A query is converted to an [[glossary/embedding]], which is compared against a vector index of pre-embedded document chunks using [[glossary/vector-similarity]] (cosine similarity or dot product). The top-k most relevant chunks are inserted into the prompt as context. The model generates a response grounded in those chunks rather than relying solely on what it memorized during training. This reduces [[glossary/hallucination]] on domain-specific and time-sensitive information. Quality depends on retrieval precision: irrelevant chunks waste context window space and can mislead the model. A [[glossary/reranker]] is often applied after initial retrieval to re-score and trim results. Evaluation requires a [[glossary/ground-truth]] set of query-answer pairs and metrics like faithfulness (does the answer match the retrieved context) and answer relevance.

## When it applies

Use RAG when: the knowledge domain changes faster than model retraining cycles allow, the corpus is too large to fit in a context window, or you need the model to cite sources. Do not use RAG as a substitute for fine-tuning when the task requires behavioral change (tone, output format) rather than knowledge injection.

## Example

A legal research assistant embeds 50,000 case summaries. A user asks about a 2024 ruling. The query vector matches the relevant case chunk. The LLM receives: `[Context: case summary text]` + `[Question: ...]` and generates a cited answer. Without RAG, the model's training cutoff would prevent it from knowing the 2024 ruling.

## Related concepts

- [[ai-agents/rag]] - architecture patterns: chunking strategies, embedding models, hybrid retrieval.
- [[glossary/embedding]] - the vector representation that makes similarity search possible.
- [[glossary/vector-similarity]] - the distance metric used to find relevant chunks.
- [[glossary/reranker]] - improves retrieval precision after the initial embedding search.
- [[glossary/hallucination]] - RAG reduces but does not eliminate hallucination.
- [[glossary/eval-set]] - a test set of query-answer pairs is needed to measure RAG quality.

## Citing this term

> See [[glossary/rag|RAG]] (llmbestpractices.com/glossary/rag).

## Related

- [[ai-agents/rag]]
- [[glossary/embedding]]
- [[glossary/vector-similarity]]
- [[glossary/reranker]]
- [[glossary/ground-truth]]
- [[glossary/hallucination]]
- [[glossary/eval-set]]