# A Mechanistic Analysis of Looped Reasoning Language Models

Authors: Hugh Blayney, Álvaro Arroyo, Johan Obando-Ceron, Pablo Samuel Castro, Aaron Courville, Michael M. Bronstein, Xiaowen Dong.

Analyzes how hidden states and attention change across repeated transformer layers. Finds that recurrent blocks often repeat the stages seen in feedforward models, while input injection and normalization affect whether those stages remain stable at greater depth.

First public: 2026-04-13.
ArXiv initial submission: 2026-04-13.

[Original paper](https://arxiv.org/abs/2604.11791) · [PDF](https://arxiv.org/pdf/2604.11791) · [HTML](https://arxiv.org/html/2604.11791v1) · [Reading-list entry](https://latent-implicit-thinking.github.io/papers/2604.11791/)
[Preprint BibTeX](https://latent-implicit-thinking.github.io/papers/2604.11791/citation.bib) · [Preprint CSL-JSON](https://latent-implicit-thinking.github.io/papers/2604.11791/citation.csl.json) · [JSON record](https://latent-implicit-thinking.github.io/papers/2604.11791/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 0.11.0; literature checked through 2026-10-26.