Ouail Kitouni

Joined

May 2022

Names

Ouail Kitouni (Preferred)

Emails

****@mit.edu (Confirmed)

Personal Links

Career & Education History

PhD student

Massachusetts Institute of Technology (mit.edu)

2019 – 2025

Advisors, Relations & Conflicts

No relations added

Expertise

Mechanistic Interpretability

Present

Physical sciences applications

Present

Physics of Deep Learning

Present

Publications

Transformers Can Navigate Mazes With Multi-Step Prediction
Niklas Nolte, Ouail Kitouni, Adina Williams, Michael Rabbat, Mark Ibrahim
- Submitted to ICLR 2025
- Readers: Everyone
The Factorization Curse: Which Tokens You Predict Underlie the Reversal Curse and More
Ouail Kitouni, Niklas Nolte, Adina Williams, Michael Rabbat, Diane Bouchacourt, Mark Ibrahim
- NeurIPS 2024 poster
- Readers: Everyone
From Neurons to Neutrons: A Case Study in Interpretability
Ouail Kitouni, Niklas Nolte, Víctor Samuel Pérez-Díaz, Sokratis Trifinopoulos, Mike Williams
- ICML 2024 Poster
- Readers: Everyone
KBFormer: A Transformer-based Diffusion Model of Structured Entities with Heterogeneous Properties
Ouail Kitouni, Niklas Nolte, James Hensman, Bhaskar Mitra
- Submitted to ICLR 2024
- Readers: Everyone
NuCLR: Nuclear Co-Learned Representations
Niklas Nolte, Ouail Kitouni, Sokratis Trifinopoulos, Subhash Kantamneni, Mike Williams
- SynS & ML @ ICML2023
- Readers: Everyone
Expressive Monotonic Neural Networks
Niklas Nolte, Ouail Kitouni, Mike Williams
- Published: 01 Feb 2023, Last Modified: 14 Jan 2026
- ICLR 2023 poster
- Readers: Everyone
Towards Understanding Grokking: An Effective Theory of Representation Learning
Ziming Liu, Ouail Kitouni, Niklas Nolte, Eric J Michaud, Max Tegmark, Mike Williams
- Published: 31 Oct 2022, Last Modified: 08 Feb 2026
- NeurIPS 2022 Accept
- Readers: Everyone

Co-Authors