# Predictive Latents for Autoregressive Language Models: Possibility, Gauge, and Obstruction

Authors: Yuzhou Hong.

Studies when predicting the next hidden state preserves information needed for next-token prediction. Gives conditions and counterexamples showing why good one-step predictions need not yield reliable multi-step hidden-state rollouts.

First public: 2026-10-20.
ArXiv initial submission: 2026-10-20.

[Original paper](https://arxiv.org/abs/2610.28540) · [PDF](https://arxiv.org/pdf/2610.28540) · [HTML](https://arxiv.org/html/2610.28540v1) · [Reading-list entry](https://latent-implicit-thinking.github.io/papers/2610.28540/)
[Preprint BibTeX](https://latent-implicit-thinking.github.io/papers/2610.28540/citation.bib) · [Preprint CSL-JSON](https://latent-implicit-thinking.github.io/papers/2610.28540/citation.csl.json) · [JSON record](https://latent-implicit-thinking.github.io/papers/2610.28540/record.json)

## Approaches

- [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.