Language Models are Few-Shot Learners (GPT-3)
Tom B. Brown and coauthors submitted Language Models are Few-Shot Learners on May 28, 2020. The paper reports GPT-3, an autoregressive language model with 175 billion parameters, and evaluates task-agnostic zero-shot, one-shot and few-shot performance without gradient updates or fine-tuning during those evaluations.
Historical classification
This verified record belongs to LINEAiGE's artificial-intelligence history layer beginning in 1956.
Why it matters
A documented language-model milestone reporting GPT-3 and large-scale task-agnostic few-shot evaluation through text interaction.
People, organizations & technologies
PEOPLE
Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeffrey Wu, Clemens Winter, Christopher Hesse, Mark Chen, Eric Sigler, Mateusz Litwin, Scott Gray, Benjamin Chess, Jack Clark, Christopher Berner, Sam McCandlish, Alec Radford, Ilya Sutskever, Dario Amodei
ORGANIZATIONS
OpenAI
TECHNOLOGIES
GPT-3, Autoregressive language model, Few-shot learning
Evidence & verification
STATE
verified
CONFIDENCE
high
LAST VERIFIED
2026-09-08T12:44:57+08:00
Title, authors, May 28 2020 submission/publication milestone, GPT-3's 175-billion-parameter scale and the described zero/one/few-shot evaluation setup are supported by the original arXiv record and OpenAI's first-party milestone page. The date is attached to this documented paper milestone, not asserted as a separate model-birth date. No ancestry, influence, causal relationship or geography is inferred.
Sources
Language Models are Few-Shot Learners
arXiv
PUBLISHED
2020-05-28
Language models are few-shot learners
OpenAI
PUBLISHED
2020-05-28
Trace the chronology
Record identity
LINEAiGE IDgpt-3-2020