PeopleofSuperintelligence

Researcher

Long Ouyang

Long Ouyang co-wrote the InstructGPT paper, posted in March 2022, which showed that fine-tuning with human feedback aligns language models with user intent. In 2016 he first-authored a paper on optimal experiment design with probabilistic programs. He also co-authored Learning to summarize with human feedback in 2020 and WebGPT in 2021. In September 2022, he gave a Scale AI presentation on InstructGPT as an OpenAI research scientist. In 2023, Arize interviewed him as a research scientist at OpenAI, and in 2024 he was listed among the authors of the GPT-4o System Card.

Now at
OpenAI
Works
15
Citations
5,952
h-index
9

Publication numbers from OpenAlex, as of 15 October 2026.

Timeline

  1. 2016

    Optimal experiment design with programs

    He is first author of Practical optimal experiment design with probabilistic programs, submitted to arXiv in August 2016.

  2. 2020

    Summaries trained on human preferences

    He co-authors Learning to summarize with human feedback, published at NeurIPS in 2020.

  3. December 2021

    WebGPT for browsing answers

    He co-authors WebGPT, which fine-tunes GPT-3 to answer long-form questions by browsing the web, as an arXiv preprint in December 2021.

  4. March 2022

    The InstructGPT paper

    He is the first-listed author of InstructGPT, posted to arXiv in March 2022, which shows that fine-tuning with human feedback aligns language models with user intent.

  5. September 2022

    A Scale AI presentation

    In September 2022, he presents InstructGPT work with Aerin Kim, an engineering manager at Scale AI, as an OpenAI research scientist.

  6. May 2023

    Research scientist at OpenAI

    Arize publishes an interview with him and Ryan Lowe, research scientists at OpenAI, in May 2023.

  7. October 2024

    A GPT-4o system card

    He is listed among the authors of the GPT-4o System Card, submitted to arXiv in October 2024.

Sources and credits

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

  1. [1608.05046] Practical optimal experiment design with probabilistic programs arxiv.org. Checked 15 October 2026.
  2. Learning to summarize with human feedback proceedings.neurips.cc. Checked 15 October 2026.
  3. [2112.09332] WebGPT: Browser-assisted question-answering with human feedback arxiv.org. Checked 15 October 2026.
  4. [2203.02155] Training language models to follow instructions with human feedback arxiv.org. Checked 15 October 2026.
  5. Training language models to follow instructions with human feedback proceedings.neurips.cc. Checked 15 October 2026.
  6. OpenAI’s InstructGPT: Aligning Language Models With Human Intent - Event | Scale Events exchange.scale.com. Checked 15 October 2026.
  7. RLHF Explained: How OpenAI Aligns Language Models arize.com. Checked 15 October 2026.
  8. [2410.21276] GPT-4o System Card arxiv.org. Checked 15 October 2026.