Minhak Song

Implicit Bias of Per-sample Adam on Separable Data: Departure from the Full-batch Regime

Understanding the Performance Gap in Preference Learning: A Dichotomy of RLHF and DPO

Through the River: Understanding the Benefit of Schedule-Free Methods for Language Model Training

Understanding Sharpness Dynamics in NN Training with a Minimalist Example: The Effects of Dataset Difficulty, Depth, Stochasticity, and More

Does SGD really happen in tiny subspaces?