Mahdi Soltanolkotabi

Learning to Recall with Transformers Beyond Orthogonal Embeddings

FoNE: Precise Single-Token Number Embeddings via Fourier Features

On the Dynamics of Learning Linear Functions with Neural Networks

The Rich and the Simple: On the Implicit Bias of Adam and SGD

MediConfusion: Can you trust your AI radiologist? Probing the reliability of multimodal medical foundation models

Test-Time Training Provably Improves Transformers as In-context Learners

Emergence and Evolution of Interpretable Concepts in Diffusion Models

ProPicker: Promptable Segmentation for Particle Picking in Cryogenic Electron Tomography