Mohammad Taufeeque

prof_pic.jpg
Research Engineer FAR.AI

I am a research engineer at FAR AI working with Dr. Adam Gleave on research projects focusing on AI Safety. I am interested in scalable approaches to mechanistic interpretability and adversarial robustness.

I graduated with a bachelors in Computer Science from IIT Bombay. For my bachelors thesis, I worked with Prof. Shivaram Kalyanakrishnan to build Fianchetto, an AI agent for the game of Reconnaissance Blind Chess that won the NeurIPS 2021 competition on RBC.

I have also interned at Microsoft Research to work on incorporating phrasal rules with LLMs on-the-fly with Prof. Sunita Sarawagi and Dr. Sriram Rajamani. In the summer of 2020, I interned at the PLRI lab of Technical University of Braunschweig, Germany to work with Prof. Thomas Deserno on human fall detection.


Publications

  1. The Obfuscation Atlas: Mapping Where Honesty Emerges in RLVR with Deception Probes
    Mohammad Taufeeque, Stefan Heimersheim, Adam Gleave, and Chris Cundy
    arXiv, 2026
  2. Path Channels and Plan Extension Kernels: a Mechanistic Description of Planning in a Sokoban RNN
    In The Fourteenth International Conference on Learning Representations, 2026
    Also appeared as a Spotlight at the Mechanistic Interpretability Workshop, NeurIPS 2025
  3. Planning in a recurrent neural network that plays Sokoban
    arXiv, 2024
    Mechanistic Interpretability Workshop, ICML 2024
  4. Exploiting Novel GPT-4 APIs
    arXiv, 2023
  5. Codebook Features: Sparse and Discrete Interpretability for Neural Networks
    Alex Tamkin, Mohammad Taufeeque, and Noah Goodman
    In Forty-first International Conference on Machine Learning, 2024
  6. imitation: Clean Imitation Learning Implementations
    arXiv, 2022
  7. Fianchetto: Speed, Belief, Guile, Caution to Win at Reconnaissance Blind Chess
    Mohammad Taufeeque*, Nitish Tongia*, and Shivaram Kalyanakrishnan
    Bachelor’s Thesis, 2022
  8. The Second NeurIPS Tournament of Reconnaissance Blind Chess
    In Proceedings of the NeurIPS 2021 Competitions and Demonstrations Track, 2022
  9. Multi-camera, multi-person, and real-time fall detection using long short term memory
    Mohammad Taufeeque*, Samad Koita*, Nicolai Spicher, and Thomas M. Deserno
    In Medical Imaging 2021: Imaging Informatics for Healthcare, Research, and Applications, 2021
  10. Randomized POMDP Planning Algorithms
    Mohammad Taufeeque and Shivaram Kalyanakrishnan
    Technical Report, 2021