Menu

Calibration‑free top‑k MoE selector saves tuning
📰

Calibration‑free top‑k MoE selector saves tuning

DEV Community: machinelearning·Papers Mache·10 days ago
#A67lxEIS
Reading 0:00
15s threshold

MoE decoding now runs at full speed without any offline threshold tuning. The new top‑k channel selector discards calibrated thresholds entirely, yet keeps the same accuracy and latency advantages that experts have come to expect.

Previously, practitioners limited routed‑expert reads by calibrating a score threshold on held‑out data; Prox + thresholds were the de facto recipe for interactive decoding at batch size one. This calibration step was required to hit a target read budget but introduced variability across inputs and even caused repetition loops in models such as Gemma‑4.

Top‑k selection removes the need for any calibrated threshold.…

Continue reading — create a free account

Join HashtagPLUS to read full articles, follow hashtags, vote, and join the conversation.