publications

publications in reverse chronological order.

* = co-first author

2026

  1. Forking Fast: Efficiently Estimating Uncertainty Dynamics in Text Generation
    Eric Bigelow, Amir Zur, Satchel Grant, Tal Haklay, Can Rager, Owen Lewis, Thomas McGrath, Jack Merullo, Ekdeep Singh Lubana, and Atticus Geiger
    arXiv preprint arXiv:2608.19611, 2026
    arXiv
  2. Stories in Space: In-Context Learning Trajectories in Conceptual Belief Space
    Eric Bigelow, Raphaël Sarfati, Daniel Wurgaft, Owen Lewis, Thomas McGrath, Jack Merullo, Atticus Geiger, and Ekdeep Singh Lubana
    arXiv preprint arXiv:2605.12412, 2026
    arXiv
  3. Towards a Cognitive Science of Large Language Models
    Eric Bigelow
    Harvard University, 2026
    HTML
  4. The Shape of Beliefs: Geometry, Dynamics, and Interventions along Representation Manifolds of Language Models’ Posteriors
    Raphaël Sarfati, Eric Bigelow, Daniel Wurgaft, Siddharth Boppana, Jack Merullo, Atticus Geiger, Owen Lewis, Thomas McGrath, and Ekdeep Singh Lubana
    arXiv preprint arXiv:2602.02315, 2026
    arXiv
  5. Reasoning Theater: Disentangling Model Beliefs from Chain-of-Thought
    Siddharth Boppana, Annabel Ma, Max Loeffler, Raphaël Sarfati, Eric Bigelow, Atticus Geiger, Owen Lewis, and Jack Merullo
    International Conference on Machine Learning (ICML), 2026
    arXiv HTML
  6. Manifold Steering Reveals the Shared Geometry of Neural Network Representation and Behavior
    Daniel Wurgaft, Can Rager, Matthew Kowal, Vasudev Shyam, Sheridan Feucht, Usha Bhalla, Tal Haklay, Eric Bigelow, Raphaël Sarfati, Thomas McGrath, Owen Lewis, Jack Merullo, Noah Goodman, Thomas Fel, Atticus Geiger, and Ekdeep Singh Lubana
    arXiv preprint arXiv:2605.05115, 2026
    arXiv
  7. Integrating Simulation and Chain-of-Thought Reasoning in Multimodal-Language Models for Physical Reasoning
    YingQiao Wang, Eric Bigelow, Tomer Ullman, Yujin Tang, and Sebastian Risi
    In ICLR Workshop on World Models: Understanding, Modelling and Scaling (Tiny Paper), 2026
    HTML PDF
  8. Priors in Time: Missing Inductive Biases for Language Model Interpretability
    Ekdeep Singh Lubana, Can Rager, Sai Sumedh R. Hindupur, Valerie Costa, Greta Tuckute, Oam Patel, Sonia Krishna Murthy, Thomas Fel, Daniel Wurgaft, Eric Bigelow, Johnny Lin, Demba Ba, Martin Wattenberg, Fernanda Viegas, Melanie Weber, and Aaron Mueller
    International Conference on Learning Representations (ICLR), 2026
    arXiv
  9. Emergence of Hierarchical Emotion Organization in Large Language Models
    Maya Okawa, Bo Zhao, Eric Bigelow, Rose Yu, Tomer Ullman, Ekdeep Singh Lubana, and Hidenori Tanaka
    International Conference on Machine Learning (ICML), 2026
    arXiv

2025

  1. Forking paths in neural text generation
    Eric Bigelow, Ari Holtzman, Hidenori Tanaka, and Tomer Ullman
    International Conference on Learning Representations (ICLR), 2025
    arXiv
  2. Belief Dynamics Reveal the Dual Nature of In-Context Learning and Activation Steering
    Eric Bigelow*, Daniel Wurgaft*, YingQiao Wang, Noah Goodman, Tomer Ullman, Hidenori Tanaka, and Ekdeep Singh Lubana
    arXiv preprint arXiv:2511.00617, 2025
    arXiv
  3. People evaluate the algorithms that drive agents’ behavior
    Eric Bigelow, and Tomer Ullman
    Open Mind, 2025
    PDF
  4. Chain of Time: In-Context Physical Simulation with Image Generation Models
    YingQiao Wang*, Eric Bigelow*, Boyi Li, and Tomer Ullman
    arXiv preprint arXiv:2511.00110, 2025
    arXiv
  5. Evaluating Self-Orienting in Language and Reasoning Models
    Eric Bigelow*, Zergham Ahmed*, and Tomer Ullman
    ICML Workshop on Assessing World Models: Methods and Metrics for Evaluating Understanding, 2025
    PDF
  6. Let’s Simulate Frame-by-Frame: In-Context Physical Simulations with Vision-Language Models
    Yingqiao Wang*, Eric Bigelow*, and Tomer Ullman
    ICML Workshop on Assessing World Models: Methods and Metrics for Evaluating Understanding, 2025
    PDF
  7. Are language models aware of the road not taken? Token-level uncertainty and hidden state dynamics
    Amir Zur, Atticus Geiger, Ekdeep Singh Lubana, and Eric Bigelow
    ICML Workshop on Actionable Interpretability, 2025
    arXiv
  8. Actual or counterfactual? Asymmetric responsibility attributions in language models
    Eric Bigelow*, Yang Xiang*, Tobias Gerstenberg, Tomer Ullman, and Samuel J Gershman
    In NeurIPS Workshop on Interpreting Cognition in Deep Learning Models (CogInterp), 2025
  9. Language models assign responsibility based on actual rather than counterfactual contributions
    Yang Xiang*, Eric Bigelow*, Tobias Gerstenberg, Tomer Ullman, and Samuel J Gershman
    In Proceedings of the Annual Meeting of the Cognitive Science Society, 2025
    PDF
  10. On (Not) Seeing Your Self: Some People May Lack Third-Person Imagery
    Eric Bigelow, and Tomer Ullman
    In Society for Philosophy and Psychology, 2025

2024

  1. In-Context Learning Dynamics with Random Binary Sequences
    Eric Bigelow, Ekdeep Singh Lubana, Robert P Dick, Hidenori Tanaka, and Tomer Ullman
    International Conference on Learning Representations (ICLR), 2024
    arXiv
  2. Foundational challenges in assuring alignment and safety of large language models
    Usman Anwar, Abulhair Saparov, Javier Rando, Daniel Paleka, Miles Turpin, Peter Hase, Ekdeep Singh Lubana, Erik Jenner, Stephen Casper, Oliver Sourbut,  ..., Eric Bigelow,  ..., and David Krueger
    Transactions in Machine Learning Research (TMLR), 2024
    HTML

2023

  1. Mechanistic mode connectivity
    Ekdeep Singh Lubana, Eric Bigelow, Robert P Dick, David Krueger, and Hidenori Tanaka
    International Conference on Machine Learning (ICML), 2023
  2. Subjective Randomness and In-Context Learning
    Eric Bigelow, Ekdeep Singh Lubana, Robert P Dick, Hidenori Tanaka, and Tomer Ullman
    In NeurIPS Workshop on UniReps: Unifying Representations in Neural Models, 2023
  3. Non-commitment in mental imagery
    Eric Bigelow, John P McCoy, and Tomer Ullman
    Cognition, 2023
    HTML

2022

  1. People’s evaluation of programs that drive agents’ behavior
    Eric Bigelow, and Tomer Ullman
    In Proceedings of the Annual Meeting of the Cognitive Science Society, 2022
  2. Mechanistic Lens on Mode Connectivity
    Ekdeep Singh Lubana, Eric Bigelow, Robert Dick, David Krueger, and Hidenori Tanaka
    NeurIPS Workshop on Distribution Shifts: Connecting Methods and Applications, 2022
  3. Opening the black box: People evaluate agents based on the algorithms that drive their behavior.
    Eric Bigelow, and Tomer Ullman
    In Society for Philosophy and Psychology, 2022

2016

  1. Inferring priors in compositional cognitive models.
    Eric Bigelow, and Steven T Piantadosi
    Proceedings of the Annual Meeting of the Cognitive Science Society, 2016
    PDF
  2. A large dataset of generalization patterns in the number game
    Eric Bigelow, and Steven Piantadosi
    Journal of Open Psychology Data, 2016
    PDF
  3. Tales of two cities: Using social media to understand idiosyncratic lifestyles in distinctive metropolitan areas
    Tianran Hu, Eric Bigelow, Jiebo Luo, and Henry Kautz
    IEEE Transactions on Big Data, 2016
    arXiv

2015

  1. On the need for imagistic modeling in story understanding
    Eric Bigelow, Daniel Scarafoni, Lenhart Schubert, and Alex Wilson
    Biologically Inspired Cognitive Architectures, 2015
    PDF