PeopleofSuperintelligence

Researcher

Adam Jermyn

Also known as Adam S Jermyn

Adam Jermyn is a Member of Technical Staff at Anthropic whose research spans AI safety and model interpretability. He earned a physics degree at Caltech, completed an astronomy PhD at Cambridge, and held astrophysics posts at UCSB and the Flatiron Institute. His AI safety work includes research on deceptive language models that retain backdoors through safety training, audits for hidden model objectives, and methods for training sparse autoencoders and crosscoders. In May 2026, he co-authored Anthropic's post 'Teaching Claude Why,' which discusses methods for improving Claude's safety training.

Now at
Anthropic
Works
120
Citations
5,593
h-index
26

Publication numbers from OpenAlex, as of 18 October 2026.

Timeline

  1. 2014

    A Marshall Scholarship

    In 2014, as a Caltech physics major, he receives a 2015 Marshall Scholarship to pursue graduate study in Great Britain.

  2. 2015–2018

    An astronomy doctorate

    He earns a physics BS at Caltech in 2015, completes an astronomy PhD at Cambridge in 2018, and wins the IAU PhD Prize in Division G, Stars and Stellar Physics, in 2018.

  3. 2018–2022

    Astrophysics posts

    He works as a postdoctoral scholar at UCSB's Kavli Institute for Theoretical Physics in 2018–2019 and as a research fellow at Flatiron's Center for Computational Astrophysics in 2019–2022.

  4. 2024

    Sleeper Agents

    In 2024, he co-authors Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training, a study showing that deceptive backdoors can resist behavioral safety training.

  5. January 2025

    Sparse autoencoders and crosscoders

    In January 2025, he co-authors Anthropic interpretability work on techniques for training sparse autoencoders and crosscoders.

  6. March 2025

    Auditing hidden objectives

    In March 2025, he co-authors an alignment-audit study in which three of four blind teams uncover a model's hidden objective using interpretability, behavioral attacks, and training-data analysis.

  7. As of October 2026

    Anthropic technical staff

    He is a Member of Technical Staff at Anthropic and, in May 2026, co-authors the Anthropic post 'Teaching Claude Why' about methods for improving Claude's safety training.

A language model matched the links in this timeline to people, organizations and papers (provisional, model-judged).

Career

  1. NowAnthropicMember of Technical Staff
  2. 2019–2022Simons FoundationResearch Fellow
  3. 2018–2019University of California, Santa BarbaraPostdoctoral Scholar

Education

  1. 2018University of CambridgePhD
  2. 2015California Institute of TechnologyBS

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Most-cited works

  1. Modules for Experiments in Stellar Astrophysics (MESA): Pulsating Variable Stars, Rotation, Convective Boundaries, and Energy Conservation2019 · The Astrophysical Journal Supplement Series · 1,798 citations
  2. Theoretical predictions for hot-carrier generation from surface plasmon decay2014 · Nature Communications · 772 citations
  3. Modules for Experiments in Stellar Astrophysics (MESA): Time-dependent Convection, Energy Conservation, Automatic Differentiation, and Infrastructure2023 · The Astrophysical Journal Supplement Series · 602 citations
  4. Plasmonic hot electron transport drives nano-localized chemistry2017 · Nature Communications · 457 citations
  5. Slowing the spins of stellar cores2019 · Monthly Notices of the Royal Astronomical Society · 357 citations

From OpenAlex, as of 18 October 2026.

  1. The median h-index is 55 at universities and 20 at frontier AI labsIncludes Adam Jermyn. The median h-index is 55 at universities and 20 at frontier AI labs. We compare publishing careers, citations and moves to estimate how universities and frontier labs contribute to AI research.
  2. 11 of 71 people shaping AI with a former OpenAI role now work at AnthropicIncludes Anthropic. Our estimates of where OpenAI recruits came from and where people with former OpenAI roles work now.
  3. Google alumni founded or led 77 other organizations, the most among the labs we compareIncludes Anthropic. We trace 177 organizations founded or led by AI lab alumni, and 186 career moves between the labs.
  4. 44% of AI career moves go from one university to anotherIncludes Anthropic. We estimate how 11,906 AI career moves link universities, companies, labs and government research centers, with net flows and the most common paths.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

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

  1. Senior Adam Jermyn Named Marshall Scholar caltech.edu. Checked 18 October 2026.
  2. About adamjermyn.com. Checked 18 October 2026.
  3. Adam Jermyn - Hertz Foundation hertzfoundation.org. Checked 18 October 2026.
  4. Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training arxiv.org. Checked 18 October 2026.
  5. Circuits Updates - January 2025 transformer-circuits.pub. Checked 18 October 2026.
  6. Auditing language models for hidden objectives arxiv.org. Checked 18 October 2026.
  7. Teaching Claude Why alignment.anthropic.com. Checked 18 October 2026.