---
title: "Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks"
description: "Retrieval-Augmented Generation gives a language model an explicit, updateable memory outside its parameters. Read it to see how a neural retriever and a sequence generator work together, and why grounding generation in retrieved evidence can improve factual, knowledge-intensive answers."
canonical_url: "https://fanout.sh/daily/2026-07-26-retrieval-augmented-generation"
md_url: "https://fanout.sh/daily/2026-07-26-retrieval-augmented-generation.md"
last_updated: "2026-07-26"
access: "public"
---

# Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks

Retrieval-Augmented Generation gives a language model an explicit, updateable memory outside its parameters. Read it to see how a neural retriever and a sequence generator work together, and why grounding generation in retrieved evidence can improve factual, knowledge-intensive answers.

## Paper details

- Authors: Lewis et al.

- Venue: NeurIPS 2020

- Track: AI Research

- Difficulty: Easy

- Reading time: 40 min

- Original paper: https://arxiv.org/abs/2005.11401

## What you will learn

- How RAG combines a dense Wikipedia retriever with a pretrained sequence-to-sequence generator.

- Why parametric memory in model weights and non-parametric memory in a document index solve different parts of the task.

- How RAG-Sequence and RAG-Token differ in whether retrieved passages stay fixed for a response or can change for each token.

## Continue exploring

[Explore RAG](https://fanout.sh/knowledge-graph?node=ai-research%3Aconcept%3Arag): Open Fanout's RAG concept node and connect the paper to retrieval, vector indexes, grounding, provenance, and knowledge-intensive generation.

---
This representation contains public Fanout content only. Protected Pro lessons, account data, billing, checkout, and pricing are not included.

Browse the public content map: https://fanout.sh/sitemap.md