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 co
Prox replaces calibrated thresholds with top-k channel selection and a weight-derived reader to accelerate Mixture-of-Experts inference while preserving accuracy.
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 selector reads, because the number of channels passing it varies across inputs; Prox with thresholds calibrated for an 18% read budget actually reads only 12.9% on average across our benchmarks.…
Summary and abstract excerpt via Hugging Face Papers ↗ · arXiv.
Channel Selection for Mixture-of-Experts Decoding Needs No Calibration Data was first seen on Hugging Face · Daily papers on Oct 28, 2026. Since then it has appeared in 24 checks on 3 different days. Its best position was #13 of 30, reached on Oct 30, 01:28 UTC. It dropped off the list after Oct 30, 01:28 UTC.
Position on Hugging Face · Daily papers at every check (lower is better).
Best rank per day (UTC).
Platform metric as reported by Hugging Face.
Archived daily, weekly and monthly lists that include this entry.
Not at the moment. It was last on the list on Oct 30, 01:28 UTC, after 3 days of trending.
#13 on Hugging Face · Daily papers, on Oct 30, 01:28 UTC. Trendoday checks this list every 120 minutes.
It was first seen on Oct 28, 2026 and has been on the list on 3 days (24 checks) since.