Publications

2026

C. Lauditi, C. Pehlevan*, and B. Bordelon*,
C. Lauditi, C. Pehlevan*, and B. Bordelon*,
M. Uzun, M. Erdogan, C. Pehlevan, and A. T. Erdogan,
M. Uzun, M. Erdogan, C. Pehlevan, and A. T. Erdogan,
K. Takanami and C. Pehlevan,
“An Asymptotic Theory of Chain-of-Thought in In-Context Learning”, Advances in Neural Information Processing Systems (NeurIPS) (spotlight), 2026.
K. Takanami and C. Pehlevan,
“An Asymptotic Theory of Chain-of-Thought in In-Context Learning”, Advances in Neural Information Processing Systems (NeurIPS) (spotlight), 2026.
A. Meterez, P. A. Nair*, D. Morwani, C. Pehlevan, S. Kakade, and A. Damian,
“A Defense of the Quadratic Model”, arXiv preprint arXiv:2607.21716, 2026.
A. Meterez, P. A. Nair*, D. Morwani, C. Pehlevan, S. Kakade, and A. Damian,
“A Defense of the Quadratic Model”, arXiv preprint arXiv:2607.21716, 2026.
A. Atanasov, J. A. Zavatone-Veth, and C. Pehlevan,
A. Atanasov, J. A. Zavatone-Veth, and C. Pehlevan,
H. T. Chaudhry et al.,
“Practical Bayesian Optimization for Scientific Discovery”, ICML 2026 Workshop on Decision-Making from Offline Datasets to Online Adaptation: Black-Box Optimization to Reinforcement Learning (Oral) (Best Paper Award Runner-Up), 2026.
H. T. Chaudhry et al.,
“Practical Bayesian Optimization for Scientific Discovery”, ICML 2026 Workshop on Decision-Making from Offline Datasets to Online Adaptation: Black-Box Optimization to Reinforcement Learning (Oral) (Best Paper Award Runner-Up), 2026.
M. I. Letey, Y. M. Lu, C. Pehlevan, and J. A. Zavatone-Veth,
“Sequential Correlations Change In-Context Learning: Effective Context Length and Architectural Mismatch”, High-dimensional Learning Dynamics (HiLD) Workshop at ICML 2026, 2026.
M. I. Letey, Y. M. Lu, C. Pehlevan, and J. A. Zavatone-Veth,
“Sequential Correlations Change In-Context Learning: Effective Context Length and Architectural Mismatch”, High-dimensional Learning Dynamics (HiLD) Workshop at ICML 2026, 2026.
W. L. Tong et al.,
W. L. Tong et al.,
I. Halder, R. Dey, and C. Pehlevan,
I. Halder, R. Dey, and C. Pehlevan,
I. Lavie, C. Lauditi*, and C. Pehlevan*,
I. Lavie, C. Lauditi*, and C. Pehlevan*,
W. Qian and C. Pehlevan,
W. Qian and C. Pehlevan,
M. A. M. Osman, W. Zhang, C. Pehlevan*, and R. I. Wilson*,
“A network model of path control”, bioRxiv, 2026.
M. A. M. Osman, W. Zhang, C. Pehlevan*, and R. I. Wilson*,
“A network model of path control”, bioRxiv, 2026.
W. L. Tong, A. Lotfi, E. Abbe, J. Susskind, C. Pehlevan, and E. Malach,
“Weighted State Souping for Efficient Long-Context Retrieval”, arXiv preprint arXiv:2610.25621, 2026.
W. L. Tong, A. Lotfi, E. Abbe, J. Susskind, C. Pehlevan, and E. Malach,
“Weighted State Souping for Efficient Long-Context Retrieval”, arXiv preprint arXiv:2610.25621, 2026.
K. Takanami and C. Pehlevan,
K. Takanami and C. Pehlevan,
A. Atanasov, B. Bordelon, J. A. Zavatone-Veth, C. Paquette, and C. Pehlevan,
“Two-Point Deterministic Equivalence for Stochastic Gradient Dynamics in Linear Models”, Advances in Theoretical and Mathematical Physics, vol. 30, no. 1, p. 36, 2026.
A. Atanasov, B. Bordelon, J. A. Zavatone-Veth, C. Paquette, and C. Pehlevan,
“Two-Point Deterministic Equivalence for Stochastic Gradient Dynamics in Linear Models”, Advances in Theoretical and Mathematical Physics, vol. 30, no. 1, p. 36, 2026.
M. I. Letey, A. Rysmakhanov, Y. M. Lu, C. Pehlevan, and J. A. Zavatone-Veth,
“What can linear attention learn from nonlinear teachers in-context?”, NeurIPS 2026 Workshop on Principles of Generative Modeling (PriGM), 2026.
M. I. Letey, A. Rysmakhanov, Y. M. Lu, C. Pehlevan, and J. A. Zavatone-Veth,
“What can linear attention learn from nonlinear teachers in-context?”, NeurIPS 2026 Workshop on Principles of Generative Modeling (PriGM), 2026.
A. Rysmakhanov, M. I. Letey, Y. M. Lu, J. A. Zavatone-Veth, and C. Pehlevan,
“In-context classification: Calibration and noise in linear attention with softmax readout”, NeurIPS 2026 Workshop on Principles of Generative Modeling (PriGM), 2026.
A. Rysmakhanov, M. I. Letey, Y. M. Lu, J. A. Zavatone-Veth, and C. Pehlevan,
“In-context classification: Calibration and noise in linear attention with softmax readout”, NeurIPS 2026 Workshop on Principles of Generative Modeling (PriGM), 2026.
E. Troiani, L. Pesce, and C. Pehlevan,
“How much easy data does a hard task need? Insights from a solvable model of attention”, UniReps: 4th Edition of the Workshop on Unifying Representations in Neural Models, 2026.
E. Troiani, L. Pesce, and C. Pehlevan,
“How much easy data does a hard task need? Insights from a solvable model of attention”, UniReps: 4th Edition of the Workshop on Unifying Representations in Neural Models, 2026.
W. L. Tong, A. Lotfi, E. Abbe, J. M. Susskind, C. Pehlevan, and E. Malach,
“Weighted State Souping for Efficient Long-Context Retrieval”, NeurIPS 2026 Workshop: Personalized, Aligned, Long-Term Memory for AI Systems, 2026.
W. L. Tong, A. Lotfi, E. Abbe, J. M. Susskind, C. Pehlevan, and E. Malach,
“Weighted State Souping for Efficient Long-Context Retrieval”, NeurIPS 2026 Workshop: Personalized, Aligned, Long-Term Memory for AI Systems, 2026.
I. Lavie, C. Lauditi*, and C. Pehlevan*,
“When Acceleration Breaks: (Stochastic) Edges of Stability in Normalized SGD”, OPT 2026: Optimization for Machine Learning (NeurIPS 2026 Workshop) (oral), 2026.
I. Lavie, C. Lauditi*, and C. Pehlevan*,
“When Acceleration Breaks: (Stochastic) Edges of Stability in Normalized SGD”, OPT 2026: Optimization for Machine Learning (NeurIPS 2026 Workshop) (oral), 2026.
A. Meterez, P. A. Nair, D. Morwani, C. Pehlevan, S. Kakade, and A. Damian,
“A Defense of the Quadratic Model”, OPT 2026: Optimization for Machine Learning (NeurIPS 2026 Workshop) (oral), 2026.
A. Meterez, P. A. Nair, D. Morwani, C. Pehlevan, S. Kakade, and A. Damian,
“A Defense of the Quadratic Model”, OPT 2026: Optimization for Machine Learning (NeurIPS 2026 Workshop) (oral), 2026.
C. Lauditi, B. Bordelon, and C. Pehlevan,
“Transfer Learning in Infinite Width Feature Learning Networks”, International Conference on Learning Representations (ICLR), 2026.
C. Lauditi, B. Bordelon, and C. Pehlevan,
“Transfer Learning in Infinite Width Feature Learning Networks”, International Conference on Learning Representations (ICLR), 2026.
J. A. Zavatone-Veth and C. Pehlevan,
“A note on the dynamics of extended-context disordered kinetic spin models”, Journal of Physics A: Mathematical and Theoretical, vol. 59, no. 4, 2026.
J. A. Zavatone-Veth and C. Pehlevan,
“A note on the dynamics of extended-context disordered kinetic spin models”, Journal of Physics A: Mathematical and Theoretical, vol. 59, no. 4, 2026.
A. Atanasov, J. A. Zavatone-Veth, and C. Pehlevan,
“Scaling and renormalization in high-dimensional regression ”, J. Stat. Mech., vol. 2026, 2026.
A. Atanasov, J. A. Zavatone-Veth, and C. Pehlevan,
“Scaling and renormalization in high-dimensional regression ”, J. Stat. Mech., vol. 2026, 2026.
W. Qian and C. Pehlevan,
“Discovering alternative solutions beyond the simplicity bias in recurrent neural networks”, International Conference on Learning Representations (ICLR), 2026.
W. Qian and C. Pehlevan,
“Discovering alternative solutions beyond the simplicity bias in recurrent neural networks”, International Conference on Learning Representations (ICLR), 2026.
M. Letey, J. A. Zavatone-Veth, Y. M. Lu, and C. Pehlevan,
“Pretrain-Test Task Alignment Governs Generalization in In-Context Learning”, International Conference on Learning Representations (ICLR), 2026.
M. Letey, J. A. Zavatone-Veth, Y. M. Lu, and C. Pehlevan,
“Pretrain-Test Task Alignment Governs Generalization in In-Context Learning”, International Conference on Learning Representations (ICLR), 2026.
B. Bordelon, M. I. Letey, and C. Pehlevan,
“Theory of Scaling Laws for In-Context Regression: Depth, Width, Context and Time”, International Conference on Learning Representations (ICLR), 2026.
B. Bordelon, M. I. Letey, and C. Pehlevan,
“Theory of Scaling Laws for In-Context Regression: Depth, Width, Context and Time”, International Conference on Learning Representations (ICLR), 2026.
A. Meterez, D. Morwani, J. Wu, C.-A. Oncescu, C. Pehlevan, and S. Kakade,
“Seesaw: Accelerating Training by Balancing Learning Rate and Batch Size Scheduling”, International Conference on Learning Representations (ICLR), 2026.
A. Meterez, D. Morwani, J. Wu, C.-A. Oncescu, C. Pehlevan, and S. Kakade,
“Seesaw: Accelerating Training by Balancing Learning Rate and Batch Size Scheduling”, International Conference on Learning Representations (ICLR), 2026.
M. Yaghoubi et al.,
M. Yaghoubi et al.,
I. Halder and C. Pehlevan,
“Demystifying LLM-as-a-Judge: Analytically Tractable Model for Inference-Time Scaling”, International Conference on Machine Learning (ICML), 2026.
I. Halder and C. Pehlevan,
“Demystifying LLM-as-a-Judge: Analytically Tractable Model for Inference-Time Scaling”, International Conference on Machine Learning (ICML), 2026.
B. Bordelon and C. Pehlevan,
B. Bordelon and C. Pehlevan,
T. Jiang, B. Bordelon, C. Pehlevan, and B. Hanin,
“Hyperparameter Transfer with Mixture-of-Expert Layers”, International Conference on Machine Learning (ICML), 2026.
T. Jiang, B. Bordelon, C. Pehlevan, and B. Hanin,
“Hyperparameter Transfer with Mixture-of-Expert Layers”, International Conference on Machine Learning (ICML), 2026.
A. Meterez*, P. A. Nair*, D. Morwani*, C. Pehlevan, and S. Kakade,
“Anytime Pretraining: Horizon-Free Learning-Rate Schedules with Weight Averaging”, High-dimensional Learning Dynamics (HiLD) Workshop at ICML (Oral), 2026.
A. Meterez*, P. A. Nair*, D. Morwani*, C. Pehlevan, and S. Kakade,
“Anytime Pretraining: Horizon-Free Learning-Rate Schedules with Weight Averaging”, High-dimensional Learning Dynamics (HiLD) Workshop at ICML (Oral), 2026.
Y. Liu, Z. Liu, C. Pehlevan, and J. Gore,
“Universal One-third Time Scaling in Learning Peaked Distributions”, International Conference on Machine Learning (ICML), 2026.
Y. Liu, Z. Liu, C. Pehlevan, and J. Gore,
“Universal One-third Time Scaling in Learning Peaked Distributions”, International Conference on Machine Learning (ICML), 2026.
B. Wang, J. Zavatone-Veth, and C. Pehlevan,
“A Random Matrix Theory Perspective on the Consistency of Diffusion Models”, International Conference on Machine Learning (ICML) (Oral) (Outstanding Paper Honorable Mention), 2026.
B. Wang, J. Zavatone-Veth, and C. Pehlevan,
“A Random Matrix Theory Perspective on the Consistency of Diffusion Models”, International Conference on Machine Learning (ICML) (Oral) (Outstanding Paper Honorable Mention), 2026.
W. L. Tong, E. Cakar, and C. Pehlevan,
“Boule or Baguette? A Study on Task Topology, Length Generalization, and the Benefit of Reasoning Traces ”, Advances in Neural Information Processing Systems (NeurIPS), 2026.
W. L. Tong, E. Cakar, and C. Pehlevan,
“Boule or Baguette? A Study on Task Topology, Length Generalization, and the Benefit of Reasoning Traces ”, Advances in Neural Information Processing Systems (NeurIPS), 2026.
D. G. Clark*, B. Bordelon*, J. A. Zavatone-Veth*, and C. Pehlevan,
D. G. Clark*, B. Bordelon*, J. A. Zavatone-Veth*, and C. Pehlevan,
I. Halder, A. Banerjee, and C. Pehlevan,
I. Halder, A. Banerjee, and C. Pehlevan,
A. Lee*, G. Kumar*, B. Bordelon, and C. Pehlevan,
“CompleteP for RL: Maintaining Feature Learning When Scaling Deep Reinforcement Learning”, International Conference on Machine Learning (ICML), 2026.
A. Lee*, G. Kumar*, B. Bordelon, and C. Pehlevan,
“CompleteP for RL: Maintaining Feature Learning When Scaling Deep Reinforcement Learning”, International Conference on Machine Learning (ICML), 2026.

2025

J. A. Zavatone-Veth, S. Yang*, J. A. Rubinfien*, and C. Pehlevan,
“How does training shape the Riemannian geometry of neural network representations?”, NeurIPS 2025 Workshop on Symmetry and Geometry in Neural Representations (NeurReps) Proceedings (Oral), 2025.
J. A. Zavatone-Veth, S. Yang*, J. A. Rubinfien*, and C. Pehlevan,
“How does training shape the Riemannian geometry of neural network representations?”, NeurIPS 2025 Workshop on Symmetry and Geometry in Neural Representations (NeurReps) Proceedings (Oral), 2025.
S. Yang, P. Liu, and C. Pehlevan,
“Convex Relaxation for Solving Large-Margin Classifiers in Hyperbolic Space ”, Transactions on Machine Learning Research, 2025.
S. Yang, P. Liu, and C. Pehlevan,
“Convex Relaxation for Solving Large-Margin Classifiers in Hyperbolic Space ”, Transactions on Machine Learning Research, 2025.
S. Yang, J. A. Zavatone-Veth, and C. Pehlevan,
“Spectral regularization for adversarially-robust representation learning”, 2025 Asilomar Conference on Signals, Systems, and Computers, 2025.
S. Yang, J. A. Zavatone-Veth, and C. Pehlevan,
“Spectral regularization for adversarially-robust representation learning”, 2025 Asilomar Conference on Signals, Systems, and Computers, 2025.
A. Atanasov*, A. Meterez*, J. B. Simon*, and C. Pehlevan,
“The Optimization Landscape of SGD Across the Feature Learning Strength”, International Conference on Learning Representations (ICLR), 2025.
A. Atanasov*, A. Meterez*, J. B. Simon*, and C. Pehlevan,
“The Optimization Landscape of SGD Across the Feature Learning Strength”, International Conference on Learning Representations (ICLR), 2025.
E. Attias, C. Pehlevan, and D. Obeid,
E. Attias, C. Pehlevan, and D. Obeid,
W. Tong and C. Pehlevan,
“MLPs Learn In-Context on Regression and Classification Tasks”, International Conference on Learning Representations (ICLR), 2025.
W. Tong and C. Pehlevan,
“MLPs Learn In-Context on Regression and Classification Tasks”, International Conference on Learning Representations (ICLR), 2025.
H. Cui, C. Pehlevan, and Y. M. Lu,
“A solvable model of learning generative diffusion: theory and insights”, Advances in Neural Information Processing Systems (NeurIPS), 2025, 2025.
H. Cui, C. Pehlevan, and Y. M. Lu,
“A solvable model of learning generative diffusion: theory and insights”, Advances in Neural Information Processing Systems (NeurIPS), 2025, 2025.
B. Bordelon and C. Pehlevan,
B. Bordelon and C. Pehlevan,
C. Lauditi, B. Bordelon, and C. Pehlevan,
“Adaptive kernel predictors from feature-learning infinite limits of neural networks”, International Conference on Machine Learning (ICML), 2025.
C. Lauditi, B. Bordelon, and C. Pehlevan,
“Adaptive kernel predictors from feature-learning infinite limits of neural networks”, International Conference on Machine Learning (ICML), 2025.
H. T. Chaudhry, M. Kulkarni, and C. Pehlevan,
“Test-time scaling meets associative memory: Challenges in subquadratic models”, ICLR Workshop: New Frontiers in Associative Memories, 2025.
H. T. Chaudhry, M. Kulkarni, and C. Pehlevan,
“Test-time scaling meets associative memory: Challenges in subquadratic models”, ICLR Workshop: New Frontiers in Associative Memories, 2025.