Chase Goddard — accepted AI-conference papers

2026

Entropic Confinement and Mode Connectivity in Overparameterized Neural Networks

Luca Di Carlo, Chase Goddard, David J. Schwab

ICLR 2026poster

Modern neural networks exhibit a striking property: basins of attraction in the loss landscape are often connected by low-loss paths, yet optimization dynamics generally remain confined to a single convex basin and rarely explore intermediate points. We resolve this paradox by identifying entropic b…

Cited by 0 Source ↗Scholar ↗

2025

When can in-context learning generalize out of task distribution?

Chase Goddard, Lindsay M. Smith, Vudtiwat Ngampruetikorn, David J. Schwab

ICML 2025poster

In-context learning (ICL) is a remarkable capability of pretrained transformers that allows models to generalize to unseen tasks after seeing only a few examples. We investigate empirically the conditions necessary on the pretraining distribution for ICL to emerge and generalize \emph{out-of-distrib…

Cited by 0 Source ↗PDF ↗Scholar ↗