Papers

How Does RL Post-training Induce Skill Composition? A Case Study on Countdown

Simon Park*, Simran Kaur*, Anirudh Goyal, Sanjeev Arora

Instruct-SkillMix: A Powerful Pipeline for LLM Instruction Tuning

Simran Kaur*, Simon Park*, Anirudh Goyal, Sanjeev Arora

Can Models Learn Skill Composition from Examples?

Haoyu Zhao, Simran Kaur, Dingli Yu, Anirudh Goyal, Sanjeev Arora

Skill-Mix: a Flexible and Expandable Family of Evaluations for AI models

Dingli Yu, Simran Kaur, Arushi Gupta, Jonah Brown-Cohen, Anirudh Goyal, Sanjeev Arora

Disentangling the Mechanisms Behind Implicit Regularization in SGD

Zachary Novack, Simran Kaur, Tanya Marwah, Saurabh Garg, Zachary Lipton

On the Maximum Hessian Eigenvalue and Generalization

Simran Kaur, Jeremy Cohen, Zachary Lipton

Gradient Descent on Neural Networks Typically Occurs at the Edge of Stability

We empirically demonstrate that full-batch gradient descent on neural network training objectives typically operates in a regime we …

Jeremy Cohen, Simran Kaur, Yuanzhi Li, Zico Kolter, Ameet Talwalker

Are Perceptually-Aligned Gradients a General Property of Robust Classifiers?

Simran Kaur, Jeremy Cohen, Zachary Lipton