Reza Bayat

Researcher, Mila - Quebec Artificial Intelligence Institute

Joined

May 2021

Names

Reza Bayat (Preferred)

Emails

****@mila.quebec (Confirmed)

****@gmail.com (Confirmed)

****@umontreal.ca (Confirmed)

Personal Links

Career & Education History

Researcher

Mila - Quebec Artificial Intelligence Institute (mila.quebec)

2024 – Present

MS student

Université de Montréal (umontreal.ca)

2021 – 2024

Intern

Mila (mila.quebec)

2020 – 2021

Advisors, Relations & Conflicts

No relations added

Expertise

memorization

generalization

mechanistic interpretability

llm reasoning

2023 – 2025

Publications

Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Thinking
Sangmin Bae, Yujin Kim, Reza Bayat, Sungnyun Kim, Jiyoun Ha, Tal Schuster, Adam Fisch, Hrayr Harutyunyan, Ziwei Ji, Aaron Courville, Se-Young Yun
- ES-FoMo III
- Readers: Everyone
Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computation
Sangmin Bae, Yujin Kim, Reza Bayat, Sungnyun Kim, Jiyoun Ha, Tal Schuster, Adam Fisch, Hrayr Harutyunyan, Ziwei Ji, Aaron Courville, Se-Young Yun
- NeurIPS 2025 poster
- Readers: Everyone
Steering Large Language Model Activations in Sparse Spaces
Reza Bayat, Ali Rahimi-Kalahroudi, Mohammad Pezeshki, Sarath Chandar, Pascal Vincent
- COLM 2025
- Readers: Everyone
The Pitfalls of Memorization: When Memorization Hurts Generalization
Reza Bayat, Mohammad Pezeshki, Elvis Dohmatob, David Lopez-Paz, Pascal Vincent
- ICLR 2025 Poster
- Readers: Everyone
Performance Control in Early Exiting to Deploy Large Models at the Same Cost of Smaller Ones
Mehrnaz Mofakhami, Reza Bayat, Ioannis Mitliagkas, Joao Monteiro, Valentina Zantedeschi
- ICLR 2025 Conference Withdrawn Submission
- Readers: Everyone
The Pitfalls of Memorization: When Memorization Hinders Generalization
Reza Bayat, Mohammad Pezeshki, Elvis Dohmatob, David Lopez-Paz, Pascal Vincent
- SciForDL Poster
- Readers: Everyone
Performance Control in Early Exiting to Deploy Large Models at the Same Cost of Smaller Ones
Mehrnaz Mofakhami, Reza Bayat, Ioannis Mitliagkas, Joao Monteiro, Valentina Zantedeschi
- ES-FoMo-II 2024 Poster
- Readers: Everyone
Towards Adversarially Robust Vision-Language Models: Insights from Design Choices and Prompt Formatting Techniques
Rishika Bhagwatkar, Shravan Nayak, Reza Bayat, Alexis Roger, Daniel Z Kaplan, Pouya Bashivan, Irina Rish
- ICML 2024 TiFA Workshop
- Readers: Everyone
Towards Adversarially Robust Vision-Language Models: Insights from Design Choices and Prompt Formatting Techniques
Rishika Bhagwatkar, Shravan Nayak, Reza Bayat, Alexis Roger, Daniel Z Kaplan, Pouya Bashivan, Irina Rish
- NextGenAISafety 2024 Poster
- Readers: Everyone
Adversarial Training with Synthesized Data: A Path to Robust and Generalizable Neural Networks
Reza Bayat, Irina Rish
- NextGenAISafety 2024 Poster
- Readers: Everyone

View all 14 publications

Co-Authors

View all 32 co-authors