Publications

Please see my Google Scholar for a complete list of papers. (^* denotes equal contribution.)

Latent Concept Disentanglement in Transformer-based Language Models

Guanzhe Hong^*, Bhavya Vasudeva^*, Vatsal Sharan, Cyrus Rashtchian, Prabhakar Raghavan, Rina Panigrahy

Submitted

In-Context Occam’s Razor: How Transformers Prefer Simpler Hypotheses on the Fly

Puneesh Deora, Bhavya Vasudeva, Tina Behnia, Christos Thrampoulidis

COLM 2025; MOSS Workshop at ICML 2025 (Oral)

The Rich and the Simple: On the Implicit Bias of Adam and SGD

Bhavya Vasudeva, Jung Whan Lee, Vatsal Sharan, Mahdi Soltanolkotabi

Submitted

Transformers Learn Low-Sensitivity Functions: Investigations and Implications

Bhavya Vasudeva^*, Deqing Fu^*, Tianyi Zhou, Elliott Kau, Youqi Huang, Vatsal Sharan

ICLR 2025

Implicit Bias and Fast Convergence Rates for Self-attention

Bhavya Vasudeva^*, Puneesh Deora^*, Christos Thrampoulidis

TMLR 2025

Mitigating Simplicity Bias in Deep Learning for Improved OOD Generalization and Robustness

Bhavya Vasudeva, Kameron Shahabi, Vatsal Sharan

TMLR 2024