Sidak Pal Singh

Cracking the Hessian: Closed-Form Hessian Spectra for Fundamental Neural Networks

Model Fusion via Neuron Interpolation

Generalized Linear Mode Connectivity for Transformers

What Does It Mean to Be a Transformer? Insights from a Theoretical Hessian Analysis

The Directionality of Optimization Trajectories in Neural Networks

Avoiding spurious sharpness minimization broadens applicability of SAM