Looped Transformers for Length Generalization

Ying Fan, Yilun Du, K. Ramchandran, Kangwook Lee
2/12/2026

No official code found. Best implementation: UW-Madison-Lee-Lab/looped-tf (14 stars, last commit 2025-04-30).

Semantic Scholar

Code Implementations(1)

UW-Madison-Lee-Lab/looped-tfCommunity53%

Code for Looped Transformers for Length Generalization (https://arxiv.org/abs/2409.15647), ICLR 2025

142C, C++Jul 30, 20241 years ago

Similar Papers

Budgeted Training for Vision Transformer

100% match2026

Dynamic Rescaling for Training GNNs

100% match1 repo2026

Universal Rates for Active Learning

100% match2026

Generalization Analysis for Controllable Learning

100% match2026

Learning Dictionary for Visual Attention

100% match2026

Unified Screening for Multiple Diseases

100% match1 repo2026

Cite this paper

@article{fan2026looped,
  title  = {Looped Transformers for Length Generalization},
  author = {Ying Fan and Yilun Du and K. Ramchandran and Kangwook Lee},
  year   = {2026},
  url    = {https://api.semanticscholar.org/CorpusID:348a2796a09c0fe3c11472b707a47eb3341dc5d9},
  journal = {ICLR 2025 2025}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/348a2796a09c0fe3c11472b707a47eb3341dc5d9.svg)](https://codeofpaper.com/paper/348a2796a09c0fe3c11472b707a47eb3341dc5d9)

Discussion