PeopleofSuperintelligence

Researcher

Jeffrey Wu

Also known as Jeff Wu

Jeffrey Wu is a researcher and engineer who co-wrote a series of language-model papers between 2019 and 2024. They co-wrote the 2019 paper Language Models are Unsupervised Multitask Learners, which introduced GPT-2. In 2020 they co-wrote Generative Pretraining From Pixels, a paper in the ICML proceedings, and Language Models are Few-Shot Learners, a NeurIPS Outstanding Paper that introduced GPT-3. Later they co-wrote WebGPT in 2021, a 2022 paper on training language models to follow instructions with human feedback, and Weak-to-Strong Generalization in 2023. In 2024 they co-wrote Scaling and evaluating sparse autoencoders, and they describe themselves as a researcher and engineer living in the Bay Area.

Timeline

  1. 2019

    The GPT-2 paper

    They co-author the 2019 paper Language Models are Unsupervised Multitask Learners, which introduces GPT-2.

  2. 2020

    Pretraining from pixels

    They co-author Generative Pretraining From Pixels, a paper in the 2020 ICML proceedings.

  3. 2020

    Few-shot learning at scale

    They co-author Language Models are Few-Shot Learners, which introduces GPT-3 and wins a NeurIPS Outstanding Paper award in 2020.

  4. December 2021

    WebGPT with human feedback

    In December 2021 they co-author WebGPT, a paper on browser-assisted question answering with human feedback.

  5. March 2022

    Following instructions with feedback

    In March 2022 they co-author Training language models to follow instructions with human feedback.

  6. December 2023

    Weak supervision for strong models

    In December 2023 they co-author Weak-to-Strong Generalization, a paper on eliciting strong capabilities with weak supervision.

  7. June 2024

    Scaling sparse autoencoders

    Their co-authored paper Scaling and evaluating sparse autoencoders appears in June 2024.

  8. As of October 2026

    Researcher and engineer

    They describe themselves as a researcher and engineer living in the Bay Area.

Sources and credits

Photo: We have no published photo of Jeffrey Wu yet. Send us a free-licensed photo.

  1. Language Models are Unsupervised Multitask Learners / APICe apice.unibo.it. Checked 14 October 2026.
  2. AI Paper Review: Language Models are Unsupervised Multitask Learners (GPT-2) freecodecamp.org. Checked 14 October 2026.
  3. Generative Pretraining From Pixels proceedings.mlr.press. Checked 14 October 2026.
  4. NeurIPS 2020 Awards neurips.cc. Checked 14 October 2026.
  5. [2005.14165] Language Models are Few-Shot Learners arxiv.org. Checked 14 October 2026.
  6. Language Models are Few-Shot Learners proceedings.neurips.cc. Checked 14 October 2026.
  7. [2112.09332v3] WebGPT: Browser-assisted question-answering with human feedback arxiv.org. Checked 14 October 2026.
  8. Paper page - WebGPT: Browser-assisted question-answering with human feedback huggingface.co. Checked 14 October 2026.
  9. [2203.02155v1] Training language models to follow instructions with human feedback arxiv.org. Checked 14 October 2026.
  10. Paper page - Training language models to follow instructions with human feedback huggingface.co. Checked 14 October 2026.
  11. [2312.09390] Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision arxiv.org. Checked 14 October 2026.
  12. [2406.04093] Scaling and evaluating sparse autoencoders arxiv.org. Checked 14 October 2026.
  13. Jeff Wu wuthejeff.com. Checked 14 October 2026.