[2026-03-09] Scaling Machine Learning Interatomic Potentials with Mixtures of Experts — Papers — Materialist