# When Does Adaptive Halting Win? Selection Rules Matter in Looped Language Models

Authors: Mahesh Reddy Pagadala.

Compares learned and untrained stopping rules on a frozen looped language model. Changing how the stopping threshold is chosen reverses which method saves more loops. Finds no demonstrated advantage for a trained probe over a simple KL-based rule; savings are estimated from stored loop traces.

First public: 2026-11-02.
ArXiv initial submission: 2026-11-02.

[Original paper](https://arxiv.org/abs/2611.02333) · [PDF](https://arxiv.org/pdf/2611.02333) · [HTML](https://arxiv.org/html/2611.02333v1) · [Reading-list entry](https://latent-implicit-thinking.github.io/papers/2611.02333/)
[Preprint BibTeX](https://latent-implicit-thinking.github.io/papers/2611.02333/citation.bib) · [Preprint CSL-JSON](https://latent-implicit-thinking.github.io/papers/2611.02333/citation.csl.json) · [JSON record](https://latent-implicit-thinking.github.io/papers/2611.02333/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.