How LLMs and coding tools work9 min read

Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks

Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, et al. · 2020 · NeurIPS 2020

Don't ask the model to remember your facts — fetch the relevant documents at question time and put them in the prompt. This is the paper that named the pattern.

Read the paper ↗

The short version

→The model's weights are parametric memory: everything it absorbed in training, frozen and unattributable. A search index is non-parametric memory: it can be updated, inspected and cited.
→RAG combines the two — a retriever pulls passages from a dense vector index of Wikipedia, and a sequence-to-sequence generator writes the answer conditioned on them.
→It beat parametric-only models on open-domain question answering and produced more specific and more factual language.
→Update your knowledge by re-indexing a document, not by retraining a model. That is the whole commercial argument for RAG.

Subscription

The Library

Everything running a software business asks of you and none of it is code: onboarding that activates, churn you can actually see coming, marketing with no budget, SEO that finds buyers, and the AI-plus-Ahrefs workflow for writing the blog that feeds it. One new playbook a week, plus the whole archive from day one.

$9/ month · cancel any time

✓A new playbook every week — onboarding, churn, growth, SEO, content
✓The full back catalogue the moment you join, not drip-fed
✓Vendor explainers: Stripe, Supabase, Resend, Vercel, the AI APIs
✓The research shelf — real papers, decoded, on learning and on LLMs
✓One email when a new one lands, off by default in one click
✓Cancel any time; you keep reading to the end of the month

first — it takes ten seconds.

No card is asked for yet — the subscription is recorded and really opens your account. See all plans →