AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Stronger Normalization-Free Transformers

Mingzhi Chen, Taiming Lu, Jiachen Zhu, Mingjie Sun, Zhuang Liu

CVPR 2026 · Architectures and efficiency › Attention and transformers · approach: Transformer expressivity and learnability · 11 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
replace LayerNorm with DerfViT-B · ImageNet-1Ktop-1 accuracy: 82.3% -> 82.8%Improved
replace LayerNorm with DerfDiT-XL/2 · ImageNet-1K generationFID: 19.94 -> 18.92Improved
replace LayerNorm with DyTDiT-XL/2 · ImageNet-1K generationFID: 19.94 -> 20.83Hurt
replace LayerNorm with DerfHyenaDNA · GenomicBenchmarksaccuracy: 85.2% -> 85.7%Improved
replace LayerNorm with Derfwav2vec 2.0 Base · LibriSpeechvalidation loss: 1.95 -> 1.93Improved
replace LayerNorm with DerfGPT-2 124M · OpenWebTextvalidation loss: 2.94 -> 2.94No clear effect
replace LayerNorm with DerfViT-B · ImageNet-1K train set (eval mode)training loss: 0.2623 -> 0.2681Hurt
add learnable shift s to erf (ablation)Derf · ImageNet-1K ViT-Btop-1 accuracy: 82.6% -> 82.8%Improved
use per-channel vector s instead of scalar (ablation)Derf · ImageNet-1K ViT-Btop-1 accuracy: 82.8% vs 82.8%No clear effect
scale tanh by 1.205 to approximate erf (ablation)DyT · ImageNet-1K ViT-Btop-1 accuracy: 82.6% -> 82.7%No clear effect
shift function off zero center (|shift| >= 2 fails) (ablation)point-wise functions · ImageNet-1K ViT-Btop-1 accuracy: 82.5% -> 81.6% (erf vertical +1)Hurt
make function unbounded (mix in linear term) (ablation)erf, tanh · ImageNet-1K ViT-Btop-1 accuracy: 82.6% -> 82.4% (erf, 0.01 linear)Hurt

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