# Inside the Loop: A Mechanistic Study of Weight-Tied Transformers on Depth-Bound Algorithmic Tasks

Authors: Magnus Sesodia, Juil Sock, Philip Torr, Christian Schroeder de Witt.

Trains looped and untied transformers to follow paths through a shuffled permutation graph without writing intermediate hops. Uses probes and activation patching to test when intermediate nodes become available and whether later computation uses them.

First public: 2026-10-16.
ArXiv initial submission: 2026-10-16.

[Original paper](https://arxiv.org/abs/2610.24135) · [PDF](https://arxiv.org/pdf/2610.24135) · [HTML](https://arxiv.org/html/2610.24135v1) · [Reading-list entry](https://latent-implicit-thinking.github.io/papers/2610.24135/)
[Preprint BibTeX](https://latent-implicit-thinking.github.io/papers/2610.24135/citation.bib) · [Preprint CSL-JSON](https://latent-implicit-thinking.github.io/papers/2610.24135/citation.csl.json) · [JSON record](https://latent-implicit-thinking.github.io/papers/2610.24135/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.