Simran Kaur

14 papers A* 5Journal 8Unranked 1
YearRankTypeTitle / Venue / Authors
2025 J jnl
How Does RL Post-training Induce Skill Composition? A Case Study on Countdown.
CoRR
Simon Park, Simran Kaur, Sanjeev Arora
2025 A* conf
Instruct-SkillMix: A Powerful Pipeline for LLM Instruction Tuning.
ICLR
Simran Kaur, Simon Park, Anirudh Goyal, Sanjeev Arora
2024 A* conf
Can Models Learn Skill Composition from Examples?
NeurIPS
Haoyu Zhao, Simran Kaur, Dingli Yu, Anirudh Goyal, Sanjeev Arora
2024 J jnl
Can Models Learn Skill Composition from Examples?
CoRR
Haoyu Zhao, Simran Kaur, Dingli Yu, Anirudh Goyal, Sanjeev Arora
2024 J jnl
Instruct-SkillMix: A Powerful Pipeline for LLM Instruction Tuning.
CoRR
Simran Kaur, Simon Park, Anirudh Goyal, Sanjeev Arora
2024 A* conf
SKILL-MIX: a Flexible and Expandable Family of Evaluations for AI Models.
ICLR
Dingli Yu, Simran Kaur, Arushi Gupta, Jonah Brown-Cohen, Anirudh Goyal, Sanjeev Arora
2023 A* conf
Disentangling the Mechanisms Behind Implicit Regularization in SGD.
ICLR
Zachary Novack, Simran Kaur, Tanya Marwah, Saurabh Garg, Zachary Chase Lipton
2023 J jnl
Skill-Mix: a Flexible and Expandable Family of Evaluations for AI models.
CoRR
Dingli Yu, Simran Kaur, Arushi Gupta, Jonah Brown-Cohen, Anirudh Goyal, Sanjeev Arora
2022 J jnl
Disentangling the Mechanisms Behind Implicit Regularization in SGD.
CoRR
Zachary Novack, Simran Kaur, Tanya Marwah, Saurabh Garg, Zachary C. Lipton
2022 — conf
On the Maximum Hessian Eigenvalue and Generalization.
ICBINB
Simran Kaur, Jeremy Cohen, Zachary Chase Lipton
2022 J jnl
On the Maximum Hessian Eigenvalue and Generalization.
CoRR
Simran Kaur, Jeremy Cohen, Zachary C. Lipton
2021 A* conf
Gradient Descent on Neural Networks Typically Occurs at the Edge of Stability.
ICLR
Jeremy Cohen, Simran Kaur, Yuanzhi Li, J. Zico Kolter, Ameet Talwalkar
2021 J jnl
Gradient Descent on Neural Networks Typically Occurs at the Edge of Stability.
CoRR
Jeremy Cohen, Simran Kaur, Yuanzhi Li, J. Zico Kolter, Ameet Talwalkar
2019 J jnl
Are Perceptually-Aligned Gradients a General Property of Robust Classifiers?
CoRR
Simran Kaur, Jeremy Cohen, Zachary C. Lipton