arXiv Machine Learning By Huiyuan Tian, Bonan Xu, Shijian Li

Beyond Geometric Complementarity: Coherent Overlap in Sparse Mixture-of-Experts Routing

Read the original on arXiv Machine Learning →

arXiv:2607. 28308v1 Announce Type: new Abstract: Sparse mixture-of-experts (MoE) language models route each token to multiple experts, suggesting a geometric account of their benefit: co-selected experts should contribute distinct representation directions.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv Machine Learning.