PeopleofSuperintelligence

Researcher

Jason Wei

Jason Wei is a researcher at Meta Superintelligence Labs. He worked at Google Brain from an October 2020 residency and at OpenAI from 2023 to 2025. His shared research includes simple text augmentation, instruction tuning, and chain-of-thought prompting. The prompting paper tested examples that showed intermediate reasoning steps on defined reasoning tasks. He also co-authored a study of abilities the authors called emergent as models grew. Later work included SimpleQA and BrowseComp, focused tests for short factual answers and difficult web searches. These projects connect methods for eliciting model behavior with checks on the results.

Works
34
Citations
6,677
h-index
20

Publication numbers from OpenAlex, as of 14 October 2026.

Photo of Jason Wei

Timeline

  1. 2019

    Expanding small text datasets

    Wei and Kai Zou published Easy Data Augmentation in 2019.

  2. October 2020

    Google Brain residency

    Wei joined Google Brain as an AI resident in October 2020.

  3. 2021

    Instruction tuning

    Wei was first author of FLAN’s instruction tuning paper in 2021.

  4. 2022

    Chain-of-thought prompting

    Wei was first author of chain-of-thought prompting in 2022.

  5. 2024–2025

    Checking factuality and browsing

    Wei co-authored SimpleQA in 2024 and BrowseComp in 2025.

  6. As of October 2026

    Meta Superintelligence Labs

    Wei works at Meta Superintelligence Labs.

Sources and credits

Photo: Jason Wei; image supplied by www.jasonwei.net. License: Not stated by source. Source: jasonwei.net. Cropped to a square and resized. Ask us to remove or replace this photo.

  1. Jason Wei: personal site jasonwei.net. Checked 11 October 2026.
  2. Jason Wei: dated CV jasonwei.net. Checked 11 October 2026.
  3. EDA: Easy Data Augmentation Techniques for Boosting Performance on Text Classification Tasks arxiv.org. Checked 11 October 2026.
  4. Finetuned Language Models Are Zero-Shot Learners arxiv.org. Checked 11 October 2026.
  5. Chain-of-Thought Prompting Elicits Reasoning in Large Language Models arxiv.org. Checked 11 October 2026.
  6. Emergent Abilities of Large Language Models arxiv.org. Checked 11 October 2026.
  7. Measuring Short-Form Factuality in Large Language Models arxiv.org. Checked 11 October 2026.
  8. BrowseComp: A Simple Yet Challenging Benchmark for Browsing Agents arxiv.org. Checked 11 October 2026.