Ziyue Li

Grokking in LLM Pretraining? Monitor Memorization-to-Generalization without Test

Routing Manifold Alignment Improves Generalization of Mixture-of-Experts LLMs

Your Mixture-of-Experts LLM Is Secretly an Embedding Model for Free

C3PO: Critical-Layer, Core-Expert, Collaborative Pathway Optimization for Test-Time Expert Re-Mixing

Many-Objective Multi-Solution Transport

R2-T2: Re-Routing in Test-Time for Multimodal Mixture-of-Experts