L2R: Low-Rank and Lipschitz-Controlled Routing for Mixture-of-Experts

arXiv:2601.21349v3 Announce Type: replace-cross Abstract: Mixture-of-Experts (MoE) models scale neural networks by conditionally activating a small subset of experts, where the router plays a central role in determining expert specialization and…

Read the full story at arXiv cs.AI ↗