【论文】Channel Selection for Mixture-of-Experts Decoding Needs No Calibration Data

arXiv 2610.37386 · 20 人推荐 · 附代码 AidenGeunGeun/calibration-free-moe（3 stars）。Mixture-of-Experts (MoE) models are bottlenecked by loading routed-expert weights from memory during interactive decoding at small batch sizes. Because most intermediate channels in a routed expert contribute little to any single token, channel selectors predict which channels matter and read only their weights, typically keeping channels above a threshold calibrated on offline data. We show that channel selection needs no calibration data. First, a calibrated threshold does not fix how much a s