Showing 2026Show all
2 papers · 1 filter
cs.LG2026
Structure Before Collapse: Transient semantic geometry in next-token prediction
Yize Zhao, Isabel Papadimitriou, Christos Thrampoulidis
Neural Collapse predicts that balanced one-hot classification pushes model representations to be equally far from each other; a symmetric configuration that depends only on the out…
cs.LG2026
Why Loss Re-weighting Works If You Stop Early: Training Dynamics of Unconstrained Features
Yize Zhao, Christos Thrampoulidis
The application of loss reweighting in modern deep learning presents a nuanced picture. While it fails to alter the terminal learning phase in overparameterized deep neural network…