Susan Wei

A Bayesian Model Selection Criterion for Selecting Pretraining Checkpoints

Stagewise Development in Transformers and the Geometry of the Loss Landscape

Leveraging free energy in pretraining model selection for improved fine-tuning

Temperature Optimization for Bayesian Deep Learning

Interventional Fairness on Partially Known Causal Graphs: A Constrained Optimization Approach

Dynamical versus Bayesian Phase Transitions in a Toy Model of Superposition