Will Mayner

AI interpretability·consciousness·neuroscience

Summary

I'm a computational neuroscientist with a mathematics & computer science background and a decade of experience developing the mathematical formalism of integrated information theory (IIT), building scientific software (PyPhi; 150+ citations), and analyzing large-scale neural data. I'm now focusing on mechanistic interpretability and AI safety, which I believe is the most impactful use of my expertise and skills.

13 peer-reviewed publications: first-author in PLOS Computational Biology, Entropy, and eNeuro; co-author in Nature Neuroscience.

Anthropic Fellows finalist (top 130 of ~5,000 applicants).

Research Interests

Mechanistic interpretability and AI safety; machine consciousness and model welfare; neural geometry; integrated information theory (IIT); computational neuroscience

Education

Ph.D. in Neuroscience

University of Wisconsin–Madison

2016–23

Sc.B. in Mathematics–Computer Science

Brown University

2009–13

AI safety training

BlueDot Technical AI Safety Course & Project

2026

Interpretability Research

Belief manifolds, and how to steer along them

Independent · BlueDot Technical AI Safety Project

April–May 2026
  • Reproduced Sarfati et al. (2026, Goodfire) "The Shape of Beliefs". LLMs represent in-context learned posteriors as curved manifolds; geometry-aware steering along either the primal manifold (activations) or dual manifold (linear field probes) changes the model's posterior with fewer side effects than naive linear steering. I connect this work to an earlier 'geometric turn' in computational neuroscience. The writeup drew a response from the paper's first author.
  • Currently extending the method to evaluation awareness and natural-language tasks (translation / code-switching).
  • Decomposed concept-injection introspection (Gemma-3 12B, Qwen-2.5 32B) into separable components: representation (what the model encodes about an injection) and report (the prompt-dependent late-layer circuitry that surfaces it), explaining apparent conflicts across prior protocols.
  • Answered two open questions from Pearson-Vogel et al. (2026): the circuitry behind prompt-framing effects, and post-training's role in building it. Concurrent with Macar et al. (2026, Anthropic) and Lederman & Mahowald (2026).

Skills

ML & Interpretability: PyTorch, Hugging Face, TransformerLens, nnsight, repeng, SAE Lens, einops, nngeometry, bitsandbytes, scikit-learn, Weights & Biases

Programming Languages: Python, C++, R, JavaScript, LaTeX

Scientific Computing: NumPy, SciPy, Pandas, Matplotlib, Dask, Mathematica, HTCondor

Mathematical Foundations: Information theory, probability theory, causal inference, linear algebra, optimization, discrete mathematics, combinatorics

Experience

Researcher

Center for Sleep and Consciousness

University of Wisconsin–Madison

2023–present
  • Lead developer and maintainer of PyPhi, the standard open-source library for IIT research (paper)
  • Developed a refinement of IIT's intrinsic information metric to measure the tradeoff between differentiation and specification (paper)
  • Extended IIT's formalism to offer an account of perception (preprint)

Graduate Research Assistant

Center for Sleep and Consciousness

University of Wisconsin–Madison

2016–23
  • Designed and conducted large-scale two-photon calcium imaging experiments in mouse visual cortex in collaboration with the Allen Institute's OpenScope program, quantifying stimulus-evoked neurophysiological differentiation (eNeuro paper)
  • Core contributor to the development of IIT's mathematical formalism; co-lead author of its most recent formulation, IIT 4.0 (PLoS Computational Biology paper)

Associate Systems Programmer

Center for Sleep and Consciousness

University of Wisconsin–Madison

2014–16
  • Conceived and built PyPhi
  • Designed and implemented evolutionary algorithms for simulating neural network-controlled agents, analyzing their emergent dynamics using information-theoretic measures
  • Conducted theoretical research contributing to the formal foundations of IIT

Publications

Lead author

  1. Mayner, W. G. P., Marshall, W., Tononi, G. (2026). Intrinsic Cause–Effect Power: The Tradeoff between Differentiation and Specification. Entropy, 28(4), 410. https://doi.org/10.3390/e28040410
  2. Mayner, W. G. P., Juel, B. E., Tononi, G. (2024, December 31). Intrinsic Meaning, Perception, and Matching. arXiv: 2412.21111 [q-bio]. https://doi.org/10.48550/arXiv.2412.21111
  3. Albantakis, L.*, Barbosa, L.*, Findlay, G.*, Grasso, M.*, Haun, A. M.*, Marshall, W.*, Mayner, W. G. P.*, Zaeemzadeh, A.*, Boly, M., Juel, B. E., Sasai, S., Fujii, K., David, I., Hendren, J., Lang, J. P., Tononi, G. (2023). Integrated Information Theory (IIT) 4.0: Formulating the Properties of Phenomenal Existence in Physical Terms. PLOS Computational Biology, 19(10), e1011465. https://doi.org/10.1371/journal.pcbi.1011465 [*co-lead author]
  4. Mayner, W. G. P., Marshall, W., Billeh, Y. N., Gandhi, S. R., Caldejon, S., Cho, A., Griffin, F., Hancock, N., Lambert, S., Lee, E. K., Luviano, J. A., Mace, K., Nayan, C., Nguyen, T. V., North, K., Seid, S., Williford, A., Cirelli, C., Groblewski, P. A., Lecoq, J., Tononi, G., Koch, C., Arkhipov, A. (2022). Measuring Stimulus-Evoked Neurophysiological Differentiation in Distinct Populations of Neurons in Mouse Visual Cortex. eNeuro, 9(1). https://doi.org/10.1523/ENEURO.0280-21.2021
  5. Mayner, W. G. P., Marshall, W., Albantakis, L., Findlay, G., Marchman, R., Tononi, G. (2018). PyPhi: A Toolbox for Integrated Information Theory. PLoS computational biology, 14(7), e1006343. https://doi.org/10.1371/journal.pcbi.1006343

Co-author

  1. Tononi, G., Albantakis, L., Barbosa, L., Boly, M., Cirelli, C., Comolatti, R., Ellia, F., Findlay, G., Casali, A. G., Grasso, M., Haun, A. M., Hendren, J., Hoel, E., Koch, C., Maier, A., Marshall, W., Massimini, M., Mayner, W. G., Oizumi, M., Szczotka, J., Tsuchiya, N., Zaeemzadeh, A. (2025). Consciousness or Pseudo-Consciousness? A Clash of Two Paradigms. Nat Neurosci, 28(4), 694–702. https://doi.org/10.1038/s41593-025-01880-y
  2. Findlay, G., Marshall, W., Albantakis, L., David, I., Mayner, W. G., Koch, C., Tononi, G. (2025, March 3). Dissociating Artificial Intelligence from Artificial Consciousness. arXiv: 2412.04571 [cs]. https://doi.org/10.48550/arXiv.2412.04571
  3. Bugnon, T., Mayner, W. G. P., Cirelli, C., Tononi, G. (2024). Sleep and Wake in a Model of the Thalamocortical System with Martinotti Cells. European Journal of Neuroscience, 59(4), 703–736. https://doi.org/10.1111/ejn.15836
  4. Gandhi, S. R., Mayner, W. G. P., Marshall, W., Billeh, Y. N., Bennett, C., Gale, S. D., Mochizuki, C., Siegle, J. H., Olsen, S., Tononi, G., Koch, C., Arkhipov, A. (2023). A Survey of Neurophysiological Differentiation across Mouse Visual Brain Areas and Timescales. Front. Comput. Neurosci., 17. https://doi.org/10.3389/fncom.2023.1040629
  5. Marshall, W., Grasso, M., Mayner, W. G. P., Zaeemzadeh, A., Barbosa, L. S., Chastain, E., Findlay, G., Sasai, S., Albantakis, L., Tononi, G. (2023). System Integrated Information. Entropy, 25(2), 334. https://doi.org/10.3390/e25020334
  6. Tononi, G., Boly, M., Grasso, M., Hendren, J., Juel, B. E., Mayner, W. G., Marshall, W., Koch, C. (2022). IIT, Half Masked and Half Disfigured. Behavioral and Brain Sciences, 45(e60), 1–19. https://doi.org/10.1017/S0140525X21001990
  7. Gomez, J. D., Mayner, W. G. P., Beheler-Amass, M., Tononi, G., Albantakis, L. (2021). Computing Integrated Information (Φ) in Discrete Dynamical Systems with Multi-Valued Elements. Entropy, 23(1), 6. https://doi.org/10.3390/e23010006
  8. Findlay, G., Marshall, W., Albantakis, L., Mayner, W. G. P., Koch, C., Tononi, G. (2019). Dissociating Intelligence from Consciousness in Artificial Systems – Implications of Integrated Information Theory. Proceedings of the 2019 towards Conscious AI Systems Symposium, AAAI SSS19. https://ceur-ws.org/Vol-2287/short7.pdf

Presentations

2025
Dissociating Artificial Intelligence From Artificial Consciousness

28th Meeting of the Association for the Scientific Study of Consciousness (Heraklion, Greece)

2025
Integrated Information Theory: A Brief Tour & Responses to Some Critiques

Center for Psychedelic & Consciousness Research, Johns Hopkins University (virtual)

2023
Meaning, Perception, and Matching: Quantifying How the Structure of Experience Matches the Environment

Neuroscience School of Advanced Studies (Venice, Italy)

2022
Integrated Information Theory: Implementation and Applications

Neuroscience Training Program, University of Wisconsin–Madison

2021
Mind-Body Matters: Clinical Applications of the Reciprocal Interaction Between Emotion and Inflammation

Neuroscience Training Program, University of Wisconsin–Madison

2020
Measuring Area- and Layer-specific Neurophysiological Differentiation with Cellular-level Resolution in Mouse Visual Cortex

Wisconsin Institute for Sleep and Consciousness

2020
deNEST: A Declarative Frontend for Building Neural Networks and Running Simulations in NEST (with Tom Bugnon)

NEST Conference 2020 (virtual)

2019
Neurophysiological Differentiation of Responses to Ecologically-Relevant vs. Irrelevant Stimuli

23rd Meeting of the Association for the Scientific Study of Consciousness (London, Ontario, Canada)

2018
PyPhi: A Toolbox for Integrated Information Theory

22nd Meeting of the Association for the Scientific Study of Consciousness (Krakow, Poland)

2016
Matching: Measuring Structural Correspondence Between Experience and the Environment Using Genetic Algorithms

PhiFest, Wisconsin Institute for Sleep and Consciousness

Awards

2019
Prize in Neuroscience, Student Poster Competition

23rd Meeting of the Association for the Scientific Study of Consciousness (London, Ontario, Canada)

Teaching

2025
Qualia Structure IIT Summer School

Teaching Assistant

Neuroscience School of Advanced Studies (Venice, Italy)

2019
Psychedelics and the Treatment of Mental Illness

Teaching Assistant (graduate seminar)

Neuroscience Training Program, University of Wisconsin–Madison

Service

2019
Precollege Enrichment Opportunity Program for Learning Excellence

Neuroscience Instructor

University of Wisconsin–Madison

Developed and taught a curriculum on neuroscience topics to precollege students from underprivileged backgrounds

2019
Wisconsin Center for Academically Talented Youth

Guest Lecturer

University of Wisconsin–Madison

Invited lecture for high school students in the Advanced Learners Program: "What Is Consciousness?"