Publications

 
Rachit Bansal, Aston Zhang, Rishabh Tiwari, Lovish Madaan, Sai Surya Duvvuri, Fnu Devvrit, David Brandfonbrener, David Alvarez-Melis, Prajjwal Bhargava, Mihir Kale, Samy Jelassi, “Let's (not) just put things in Context: Test-time Training for Long-context LLMs”, submitted, 2025.

Parsa Mirtaheri*, Ezra Edelman*, Samy Jelassi, Eran Malach, Enric Boix-Adsera “Let Me Think! A Long Chain-of-Thought Can Be Worth Exponentially Many Short Ones” , 39th Conference on Neural Information Processing Systems (NeurIPS), 2025.
*Equal contribution

Rosie Zhao*, Alexandru Meterez*, Sham Kakade, Cengiz Pehlevan, Samy Jelassi†, and Eran Malach†, “Echo chamber: RL post-training amplifies behaviors learned in pretraining” , 2nd Conference on Language Modeling (COLM), 2025.
*Equal contribution  |  †Equal senior contribution

Tian Qin, David Alvarez-Melis†, Samy Jelassi†, and Eran Malach†, “To backtrack or not to backtrack: When sequential search limits model reasoning” , 2nd Conference on Language Modeling (COLM), 2025.
†Equal senior contribution

Noah Golowich, Samy Jelassi, David Brandfonbrener, Sham Kakade, Eran Malach, “The Role of Sparsity for Length Generalization in Transformers”, 42nd International Conference on Machine Learning (ICML), 2025.

Kaiying Hou, David Brandfonbrener, Sham Kakade, Samy Jelassi†, and Eran Malach†, “Universal Length Generalization with Turing Programs” , 42nd International Conference on Machine Learning (ICML), 2025.
†Equal senior contribution

Samy Jelassi, Clara Mohri, David Brandfonbrener, Alex Gu, Nikhil Vyas, Nikhil Anand, David Alvarez-Melis, Yuanzhi Li, Sham M. Kakade, and Eran Malach, “Mixture of Parrots: Experts improve memorization more than reasoning”, 13th International Conference on Learning Representations (ICLR), 2025. Oral presentation (top 10%) at at the “Mathematics of modern machine learning” workshop, NeurIPS 2024.

Jyothish Pari, Samy Jelassi, and Pulkit Agrawal, “Collective Model Intelligence Requires Compatible Specialization”, In submission, 2024.

Akshara Prabhakar, Yuanzhi Li, Karthik Narasimhan, Sham Kakade, Eran Malach, and Samy Jelassi, “LoRA Soups: Merging LoRAs for Practical Skill Composition Tasks”, 31st International Conference on Computational Linguistics (COLING), Industry track, 2025.

Kenneth Li, Samy Jelassi, Hugh Zhang, Sham Kakade, Martin Wattenberg, and David Brandfonbrener, “Q-Probe: A Lightweight Approach to Reward Maximization for Language Models”, 41st International Conference on Machine Learning (ICML), 2024.

Samy Jelassi, David Brandfonbrener, Sham M. Kakade, and Eran Malach, “Repeat After Me: Transformers are Better than State Space Models at Copying”, 41st International Conference on Machine Learning (ICML), 2024.

Samy Jelassi, Stéphane d'Ascoli, Carles Domingo-Enrich, Yuhuai Wu, Yuanzhi Li, and François Charton, “Length Generalization in Arithmetic Transformers”, 2023.

Samy Jelassi, Boris Hanin, Ziwei Ji, Sashank J. Reddi, Srinadh Bhojanapalli, and Sanjiv Kumar, “Depth Dependence of μP Learning Rates in ReLU MLPs”, 2023.

Samy Jelassi, Michael Sander and Yuanzhi Li, “Vision Transformers provably learn spatial structure”, 36th Conference on Neural Information Processing Systems (NeurIPS), 2022.

Samy Jelassi*, David Dobre*, Arthur Mensch, Yuanzhi Li, and Gauthier Gidel, “Dissecting adaptive methods in GANs”, 2022.

Samy Jelassi, and Yuanzhi Li, “Towards understanding how momentum improves generalization in deep learning”, 39th International Conference on Machine Learning (ICML), 2022.
Oral presentation (top 5%) at at the “Overparameterization: Pitfalls & Opportunities” workshop, ICML 2021.

Luca Venturi, Samy Jelassi, Tristan Ozuch, and Joan Bruna, “Depth separation beyond radial functions”, Journal of Machine Learning Research (JMLR), 2022.

Aaron Defazio, and Samy Jelassi, “Adaptivity without Compromise: A Momentumized, Adaptive, Dual Averaged Gradient Method for Stochastic Optimization”, Journal of Machine Learning Research (JMLR), 2022.

Jad Rahme, Samy Jelassi, and S. Matthew Weinberg, “Auction learning as a two-player game”, 9th International Conference on Learning Representations (ICLR), 2021.

Jad Rahme, Samy Jelassi, Joan Bruna, and S. Matthew Weinberg, “A Permutation-Equivariant Neural Network Architecture For Auction Design”, 35th AAAI Conference on Artificial Intelligence, 2021.

Samy Jelassi, Carles Domingo-Enrich, Damien Scieur, Arthur Mensch, Joan Bruna, “Extragradient with player sampling for faster Nash equilibrium finding”, 37th International Conference on Machine Learning (ICML) 2020.

Carles Domingo-Enrich, Samy Jelassi, Arthur Mensch, Grant Rotskoff, and Joan Bruna, “A mean-field analysis of two-player zero-sum games”, 33rd Conference on Neural Information Processing Systems (NeurIPS) 2019.

Othmane Sebbouh, Nidham Gazagnadou, Samy Jelassi, Francis Bach, and Robert M. Gower, “Towards closing the gap between the theory and practice of SVRG”, 33rd Conference on Neural Information Processing Systems (NeurIPS) 2019.

Grant Rotskoff, Samy Jelassi, Joan Bruna, and Eric Vanden-Eijnden, “Global convergence of neuron birth-death dynamics”, 36th International Conference on Machine Learning (ICML) 2019.

Thomas Pumir, Samy Jelassi, and Nicolas Boumal, “Smoothed analysis of the low-rank approach for smooth semidefinite programs”, Oral presentation (top 3%) at the 32nd Conference on Neural Information Processing Systems (NeurIPS) 2018.