pith. sign in

We run all the experiments on one NVIDIA RTX PRO 6000 Blackwell

For training on ImageNet-1K datasets, we set batch size = 1024, we set p = 0 · 2022 · arXiv 1495.6395

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
read on arXiv browse 1 citing papers

fields

cs.LG 1

years

2026 1

verdicts

UNVERDICTED 1

representative citing papers

HTMuon: Improving Muon via Heavy-Tailed Spectral Correction

cs.LG · 2026-03-10 · unverdicted · novelty 5.0

HTMuon modifies Muon to produce heavier-tailed updates and weight spectra via HT-SR theory, yielding up to 0.98 lower perplexity on LLaMA pretraining and serving as a plug-in for other Muon variants.

citing papers explorer

Showing 1 of 1 citing paper.

  • HTMuon: Improving Muon via Heavy-Tailed Spectral Correction cs.LG · 2026-03-10 · unverdicted · none · ref 39

    HTMuon modifies Muon to produce heavier-tailed updates and weight spectra via HT-SR theory, yielding up to 0.98 lower perplexity on LLaMA pretraining and serving as a plug-in for other Muon variants.