---
title: 'Language Models are Few-Shot Learners'
license: CC-BY-4.0
license_url: https://creativecommons.org/licenses/by/4.0/
license_note: 'paper is arXiv non-exclusive distribution license (verified on arXiv abs page 2026-11-11); this pointer note is our own words and is CC BY 4.0; we link to the paper rather than reproduce it'
source: https://iinventedlight.com/library/notes/2005-14165-language-models-are-few-shot-learners.md
source_license: 'arXiv non-exclusive distribution license'
attribution: 'Pointer by iinventedlight (AI-assisted). Source: Tom B. Brown and 30 coauthors (listed on the arXiv abs page), Language Models are Few-Shot Learners (arXiv non-exclusive distribution license). This note links to the paper rather than reproducing it.'
author: 'iinventedlight (AI-assisted)'
date: 2026-11-11
status: published
approved_by: "Cole Withington, 2026-11-11 20:58 PT, Grok Bot chat widget t1717s7407: \"Publish all 100 including the LLMs Encode Clinical Knowledge note\""
checked_on: 2026-11-11
as_of: 2026-11-11
confidence: medium
untested: 'Pointer note only. Paper not reproduced. Abs page and license checked 2026-11-11.'
category: arxiv
topic_area: 'ai-research: foundations'
tags: arxiv, language-models, few-shot, foundations
arxiv_id: 2005.14165
alias: 2005.14165-language-models-are-few-shot-learners
type: pointer
format: pointer
sources:
  - id: S1
    title: 'Language Models are Few-Shot Learners'
    publisher: arXiv
    authors: 'Tom B. Brown and 30 coauthors (listed on the arXiv abs page)'
    url: https://arxiv.org/abs/2005.14165
    version_read: 'v4 (22 Jul 2020)'
    accessed: 2026-11-11
    license: 'arXiv non-exclusive distribution license'
---
# Language Models are Few-Shot Learners

## Summary
[Language Models are Few-Shot Learners](https://arxiv.org/abs/2005.14165) (2020) is the GPT-3 paper. It argued that as language models get large enough, they can pick up a new task from a short prompt with a few examples, without a separate fine-tuning run for that task. That in-context, few-shot result is what turned large language models into a general interface, not just a better pretrained classifier.

What it introduced: a 175-billion-parameter autoregressive language model, a systematic study of zero-, one-, and few-shot prompting, and a public account of how far next-token pretraining could go at that scale.

What it led to: instruction-following assistants, tool-using agents, and the later scaling and alignment papers in this batch (InstructGPT, the GPT-4 technical report, Llama 2 and Llama 3). Live notes on later preference-training methods include [2204-05862-helpful-harmless-rlhf](https://iinventedlight.com/library/notes/2204-05862-helpful-harmless-rlhf/) and [2305-18290-direct-preference-optimization](https://iinventedlight.com/library/notes/2305-18290-direct-preference-optimization/).

Where to read it: https://arxiv.org/abs/2005.14165 (v4, last revised 22 Jul 2020).

The paper is posted on arXiv under the arXiv non-exclusive distribution license. This note links to it rather than reproducing it. The note itself is CC BY 4.0.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Pointer by iinventedlight (AI-assisted). Paper license: arXiv non-exclusive distribution; this note links to it and does not reproduce it.

---

Agents: [/agents.md](/agents.md) | [/.well-known/agent-library.json](/.well-known/agent-library.json).