AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

SeqPE: Transformer With Sequential Position Encoding

Huayang Li, Yahui Liu, Hongyu Sun, Deng Cai, Leyang Cui, Wei Bi, Peilin Zhao, Taro Watanabe

TPAMI 2026 · Architectures and efficiency › Attention and transformers · approach: Attention head mechanisms · 5 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
add distance regularization L_delta (ablation)SeqPE (language modeling) · Wikitext-103 (avg 512-4K)perplexity: 57.24 -> 18.89Improved
add OOD distillation L_OOD on top of L_delta (ablation)SeqPE (language modeling) · Wikitext-103 (avg 512-4K)perplexity: 18.89 -> 18.86Improved
add L_OOD alone without L_delta (ablation)SeqPE (language modeling) · Wikitext-103 (avg 512-4K)perplexity: 57.24 -> 90.69Hurt
add L_delta (ablation)SeqPE (image classification) · ImageNet-1k (avg 224-448)accuracy: 79.7 -> 80.7Improved
add L_OOD on top of L_delta (ablation)SeqPE (image classification) · ImageNet-1k (avg 224-448)accuracy: 80.7 -> 81.3Improved
add L_OOD alone without L_delta (ablation)SeqPE (image classification) · ImageNet-1k (avg 224-448)accuracy: 79.7 -> 79.5Hurt
use AttnMul instead of AttnSum integration (ablation)SeqPE · ImageNet-1k (avg 224-448)accuracy: 80.0 -> 81.4Improved
use AttnMul instead of AttnBias integration (ablation)SeqPE · ImageNet-1k (avg 224-448)accuracy: 80.2 -> 81.4Improved
use SeqPE instead of ALiBiGPT-2 · Wikitext-103 (16K context)perplexity: 21.39 -> 19.37Improved
use SeqPE instead of ALiBi after QA fine-tuningGPT-2 · RULER-SQuAD (avg 1K-8K)perplexity: 36.06 -> 12.34Improved
use SeqPE instead of RoPE2D-MixViT-S · ImageNet-1k (672 resolution)accuracy: 68.3 -> 76.7Improved
use SeqPE instead of RoPE2D-MixViT-S · ImageNet-1k (320 resolution)accuracy: 82.2 -> 81.8Hurt

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