# John Langford on "Almost Free State Prediction Separation"

> One of the two first authors announces the paper in two posts: state-prediction separation put through an optimization process, which yields free pause tokens that let a transformer predict better at almost no cost at inference, with the training overhead cut from about 700% to an amortized 10 to 20%.

- Author: John Langford ([@JohnCLangford](https://x.com/JohnCLangford))
- Posted: 2026-09-08, 2 posts
- Original: https://x.com/JohnCLangford/status/2097339417346093102
- About: [Almost Free State Prediction Separation](https://introspection.infinite.fun/papers/langford2026-almost-free-sps.md)

The post text below is quoted verbatim. Figure descriptions are written by this wiki.

## 1/2

> A quick fun paper: We ran State Prediction Separation https://arxiv.org/pdf/2607.01218 through an optimization process https://arxiv.org/abs/2609.03807 finding that it works at higher op and yielding "free pause tokens" which allow a transformer to make higher quality inference for (near) free.

[Post 1 on X](https://x.com/JohnCLangford/status/2097339417346093102)

## 2/2

> The training time overhead is reduced from ~700% in SPS to an amortized 10-20% depending on precise choices.  Overall, an improvement on basic transformers which may be widely useful.
>
> Joint with @nthngdy, Giovanni, @yoavartzi, @Real_HDong , Ying, @rosaguga, @ZhengZhan13.

[Post 2 on X](https://x.com/JohnCLangford/status/2097339419778855215)

---

Source: https://introspection.infinite.fun/threads/johnclangford-almost-free-sps · Part of the [LLM Introspection Wiki](https://introspection.infinite.fun/index.md) · Index for agents: [llms.txt](https://introspection.infinite.fun/llms.txt)