Author name: Sanghyeok Chu, Pyunghwan Ahn, Gwangmo Song, SeungHwan Kim, Honglak Lee, Bohyung Han

arXiv:2604.13508v2 Announce Type: replace
Abstract: Sparse Upcycling provides an efficient way to initialize a Mixture-of-Experts (MoE) model from pretrained dense weights instead of training from scratch. However, since all experts start from identic…

arXiv:2604.13508v1 Announce Type: new
Abstract: Sparse Upcycling provides an efficient way to initialize a Mixture-of-Experts (MoE) model from pretrained dense weights instead of training from scratch. However, since all experts start from identical w…