Explore
Stack
Learn
Search the catalog
⌘K
$
/
₹
Sign up free
Login
Back to papers
September 10, 2026
cs.AI
cs.CL
cs.LG
A Unified Per-Token Gating Family for On-Policy Distillation: FKL/RKL Mixing with Multi-Channel and Bias Coefficients
Suwan Wu
,
Yumeng Lin
,
Pengcheng Yuan
,
Xiaolong Jiang
Original Abstract
Read on arXiv
Download PDF
Categories
cs.AI, cs.CL, cs.LG
A Unified Per-Token Gating Family for | One9Founders