AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Muon Outperforms Adam in Tail-End Associative Memory Learning

Shuche Wang, Fengzhuo Zhang, Jiaxiang Li, Cunxiao Du, Chao Du, Tianyu Pang, Zhuoran Yang, Mingyi Hong, Vincent Y. F. Tan

ICLR 2026 · Architectures and efficiency › Attention and transformers · approach: Sparse attention for long-context LLMs · 48 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use Muon instead of Adam on all weights160M NanoGPT, non-gated FFN · Validation loss (10000 steps, non-gated FFN)validation loss: 3.9242 -> 3.5654Improved
add Muon on VO weights (Adam elsewhere) vs Muon on QK only (ablation)160M NanoGPT, non-gated FFN · Validation loss (10000 steps, non-gated FFN)validation loss: 3.8925 -> 3.7644Improved
add Muon on FFN to VO (VO+FFN vs VO only) (ablation)160M NanoGPT, non-gated FFN · Validation loss (10000 steps, non-gated FFN)validation loss: 3.7644 -> 3.5858Improved
add Muon on QK to VO+FFN (full Muon vs VO+FFN) (ablation)160M NanoGPT, non-gated FFN · Validation loss (10000 steps, non-gated FFN)validation loss: 3.5858 -> 3.5654Improved
add Muon on W_o to V+FFN (ablation)160M NanoGPT, non-gated FFN · Validation loss (10000 steps, non-gated FFN)validation loss: 3.6702 -> 3.5858Improved
add Muon on W_v to O+FFN (ablation)160M NanoGPT, non-gated FFN · Validation loss (10000 steps, non-gated FFN)validation loss: 3.6042 -> 3.5858Improved
use Muon on W_out instead of W_in only (ablation)160M NanoGPT, non-gated FFN · Validation loss (10000 steps, non-gated FFN)validation loss: 3.7170 -> 3.7023Improved
add Muon on W_out to VO (vs W_in) (ablation)160M NanoGPT, non-gated FFN · Validation loss (10000 steps, non-gated FFN)validation loss: 3.6778 -> 3.6054Improved
add Muon on VO+FFN instead of VO only (ablation)160M NanoGPT, gated FFN · Validation loss (10000 steps, gated FFN)validation loss: 3.6874 -> 3.5312Improved
add Muon on all weights vs VO+FFN (ablation)160M NanoGPT, gated FFN · Validation loss (10000 steps, gated FFN)validation loss: 3.5312 -> 3.5125Improved
add Muon on W_gate and W_in and W_out to VO (vs W_in+W_gate) (ablation)160M NanoGPT, gated FFN · Validation loss (10000 steps, gated FFN)validation loss: 3.5681 -> 3.5482Improved
use Muon on all weights instead of AdamHeavy-tail knowledge task, group 15 · 10000 steps, non-gated FFNfirst token accuracy: 0.264 -> 0.976Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy