# Enhancing Auto-regressive Chain-of-Thought through Loop-Aligned Reasoning

Authors: Qifan Yu, Zhenyu He, Sijie Li, Xun Zhou, Jun Zhang, Jingjing Xu, Di He.

Trains a looped transformer to predict the tokens of a reasoning step in parallel at each iteration. Uses its longer reasoning traces to train an autoregressive model, testing length generalization on arithmetic and dynamic-programming tasks.

First public: 2025-02-12.
ArXiv initial submission: 2025-02-12.

[Original paper](https://arxiv.org/abs/2502.08482) · [PDF](https://arxiv.org/pdf/2502.08482) · [HTML](https://arxiv.org/html/2502.08482v1) · [Reading-list entry](https://latent-implicit-thinking.github.io/papers/2502.08482/)
[Preprint BibTeX](https://latent-implicit-thinking.github.io/papers/2502.08482/citation.bib) · [Preprint CSL-JSON](https://latent-implicit-thinking.github.io/papers/2502.08482/citation.csl.json) · [JSON record](https://latent-implicit-thinking.github.io/papers/2502.08482/record.json)

## Approaches

- [Recurrent reasoning](https://latent-implicit-thinking.github.io/topics/recurrent-reasoning/): Repeat parts of the model or keep intermediate information in memory while working through a problem. Some methods also learn how many passes to run.
- [Alternative generation methods](https://latent-implicit-thinking.github.io/topics/alternative-reasoning-schedules/): Change how intermediate steps are generated, revised, or searched—for example, by refining several steps in parallel. Some produce written chains of thought.

Reading-list version 0.15.2; literature checked through 2026-11-03.