Signals 4 · free daily AI digest

MeqMuon: Matrix-Equilibrating Muon for LLM Pretraining

Paper recorded by Signals 4 on 2026-09-28 in cs.LG. Abstract reproduced from arXiv; link to the original below.

Published 2026-09-28 on arXiv · recorded by Signals 4 on 2026-09-29

Category: cs.LG · 机器学习 · first seen 2026-09-29

Abstract

The success of large language models (LLMs) has been accompanied by continued growth in model size and pretraining costs. Muon offers high accuracy and training efficiency in LLM pretraining. Recent work introduces row-wise normalization into Muon to balance update magnitudes and improve pretraining performance. However, row-wise normalization alone cannot accommodate different imbalance patterns

Read on arXiv →

#8 most recent of 322 cs.LG papers we have recorded · ↑ newer: ScAn-Bench: Evaluating Scaling Analysis Methodology · ↓ older: Provable Benefits of Regularization: Fast Rates for Adversarial Imitat
Cite this page: MeqMuon: Matrix-Equilibrating Muon for LLM Pretraining: the #8 most recent of 322 cs.LG papers we have recorded (as of 2026-09-28). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/meqmuon-matrix-equilibrating-muon-for-llm-pretraining.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable: papers.json
Related: More cs.LG papers · arXiv signals · All papers · Today in AI
Get 4 AI signals a day by email — free.
Subscribe free → See all plans →
Get 4 AI signals a day by email — free
All models · All repos · By company · Daily editions