# Loop Before You Leap: What Recurrent Depth Can and Cannot Buy in Data-Limited Pretraining

Authors: Daichi Hotta, Yohei Kobashi, Takeshi Kojima, Yusuke Iwasawa, Yutaka Matsuo.

Compares shared-layer recurrent transformers with standard models under limited language data, with and without image features. Separates parameter and compute comparisons and measures how data composition and visual input affect linguistic performance.

First public: 2026-11-04.
ArXiv initial submission: 2026-11-04.

[Original paper](https://arxiv.org/abs/2611.06206) · [PDF](https://arxiv.org/pdf/2611.06206) · [HTML](https://arxiv.org/html/2611.06206v1) · [Reading-list entry](https://latent-implicit-thinking.github.io/papers/2611.06206/)
[Preprint BibTeX](https://latent-implicit-thinking.github.io/papers/2611.06206/citation.bib) · [Preprint CSL-JSON](https://latent-implicit-thinking.github.io/papers/2611.06206/citation.csl.json) · [JSON record](https://latent-implicit-thinking.github.io/papers/2611.06206/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.
- [Analysis and evaluation](https://latent-implicit-thinking.github.io/topics/implicit-reasoning-evaluation/): Measure how well these methods work, investigate what their internal computation does, and identify where they fail. This includes experiments, studies of model internals, and mathematical results.

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