PeopleofSuperintelligence

Researcher

Alex Tamkin

As of October 2026, Alex Tamkin is an Anthropic researcher whose work covers language-model interpretability and AI's effects in society. At Stanford, he first-authored DABS in 2022, a domain-agnostic benchmark for self-supervised learning across different data types. With Mohammad Taufeeque, he coauthored Codebook Features in 2023, a method that gives neural networks sparse, discrete hidden states. From 2024 through 2026, he coauthored studies of Claude 3 Sonnet features, public input into model constitutions, privacy-preserving analysis of Claude conversations, economic patterns in Claude use, and AI-assisted coding skills. He holds a machine-learning PhD from Stanford, where Noah Goodman advised him.

Now at
Anthropic
Works
47
Citations
3,342
h-index
14
Badges

Publication numbers from OpenAlex, as of 18 October 2026.

Timeline

  1. 2022

    A benchmark for self-supervision

    In 2022, he is a doctoral candidate in computer science and first author of DABS, a domain-agnostic benchmark for comparing self-supervised learning across domains.

  2. October 2023

    Sparse features for interpretability

    In October 2023, he and Mohammad Taufeeque introduce Codebook Features, which trains neural networks with sparse, discrete hidden states to improve interpretability.

  3. May 2024

    Interpretable features in Claude

    In May 2024, he coauthors Scaling Monosemanticity, where sparse autoencoders extract interpretable features from Claude 3 Sonnet and researchers use those features to steer model behavior.

  4. June 2024

    A public-input AI constitution

    In June 2024, he coauthors Collective Constitutional AI, a process for using public input to draft model principles; the study fine-tunes a language model on principles sourced from the public.

  5. December 2024

    Privacy-preserving analysis of Claude

    In December 2024, he coauthors Clio, a privacy-preserving platform that analyzes millions of AI conversations and surfaces aggregated patterns in real-world use.

  6. November 2025

    Measuring Claude use across economies

    In November 2025, he coauthors an Anthropic Economic Index report that tracks Claude use across countries, U.S. states, and businesses through privacy-preserving analysis of conversations and API transcripts.

  7. January 2026

    AI assistance and coding skills

    In January 2026, he co-leads a randomized study with Judy Hanwen Shen on how AI assistance affects coding skill formation, using software developers as participants.

  8. As of October 2026

    Researcher at Anthropic

    As of October 2026, he is a researcher at Anthropic, with interests in understanding and improving AI's impacts on the world. He holds a machine-learning PhD from Stanford, where Noah Goodman advised him.

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

Career

  1. Since 2023AnthropicResearcher

Education

  1. Stanford UniversityPhD

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

Most-cited works

  1. On the Opportunities and Risks of Foundation Models2021 · 2,638 citationsWith Antoine Bosselut, Christopher D. Manning, Christopher Potts, Christopher Ré, Dan Jurafsky and Daniel E. Ho
  2. Understanding the Capabilities, Limitations, and Societal Impact of Large Language Models2021 · 134 citations
  3. Collective Constitutional AI: Aligning a Language Model with Public Input2024 · 76 citationsWith Divya Siddarth, Esin Durmus, Liane Lovitt, Saffron Huang and Thomas Liao
  4. Drone.io: A Gestural and Visual Interface for Human-Drone Interaction2019 · ACM/IEEE International Conference on Human-Robot Interaction (HRI) · 61 citations
  5. Towards Measuring the Representation of Subjective Global Opinions in Language Models2023 · 47 citationsWith Anton Bakhtin, Esin Durmus, Liane Lovitt, Thomas Liao and Zac Hatfield-Dodds

From OpenAlex, as of 18 October 2026.

  1. The median h-index is 55 at universities and 20 at frontier AI labsIncludes Alex Tamkin. 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. 13% of AI career moves in the 2020s went from a university to a companyIncludes Alex Tamkin. Our estimates of how university-to-company career transitions changed by decade, including roles added alongside an earlier job.
  3. 11 of 73 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.
  4. Google alumni founded or led 79 other organizations, the most among the labs we compareIncludes Anthropic. We trace 181 organizations founded or led by AI lab alumni, and 187 career moves between the labs.

Colleagues

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

Sources and credits

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

  1. Are Universal Self-Supervised Learning Algorithms Within Reach? crfm.stanford.edu. Checked 18 October 2026.
  2. Codebook Features: Sparse and Discrete Interpretability for Neural Networks ai.stanford.edu. Checked 18 October 2026.
  3. Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet transformer-circuits.pub. Checked 18 October 2026.
  4. Collective Constitutional AI: Aligning a Language Model with Public Input arxiv.org. Checked 18 October 2026.
  5. Clio: Privacy-Preserving Insights into Real-World AI Use arxiv.org. Checked 18 October 2026.
  6. Anthropic Economic Index report: Uneven geographic and enterprise AI adoption arxiv.org. Checked 18 October 2026.
  7. How AI assistance impacts the formation of coding skills anthropic.com. Checked 18 October 2026.
  8. Alex Tamkin alextamkin.com. Checked 18 October 2026.
  9. Alex Tamkin - Research Scientist | FAR.AI far.ai. Checked 18 October 2026.