Language Models are Few-Shot Learners ​

← Back to topic

Authors: Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, et al.
Year: 2020
Journal: NeurIPS
DOI: 10.48550/arXiv.2005.14165
Publisher: https://arxiv.org/abs/2005.14165

Keywords: gpt-3, few-shot, in-context learning

Abstract ​

We train GPT-3 an autoregressive language model with 175 billion parameters and test its performance in the few-shot setting.

Cite this paper ​

@misc{gpt32020,
  title  = {Language Models are Few-Shot Learners},
  author = {Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, et al.},
  year   = {2020},
  journal = {NeurIPS},
  doi    = {10.48550/arXiv.2005.14165},
  url    = {https://doi.org/10.48550/arXiv.2005.14165},
}

Source files ​

• metadata.json
• paper.bib