# Natural Latents: Latent Variables Stable Across Ontologies

arXiv: https://arxiv.org/abs/2509.03780 · HTML: https://mlsanity.com/paper/2509.03780/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 2 of its 5 references that are in the map.
- First posted on arXiv 4 Sep 2025.

## Builds on (most cited first, 2 of 2)

- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [The Platonic Representation Hypothesis](https://mlsanity.com/paper/2405.07987/) (2024)

## Built on by (most cited first, 2 of 2)

- [Critical Percolation as a Synthetic Data Model for Interpretability](https://mlsanity.com/paper/2606.20347/) (2026)
- [Math for AI safety: an invitation for mathematicians](https://mlsanity.com/paper/2609.15289/) (2026)

## Abstract (arXiv)

Suppose two Bayesian agents each learn a generative model of the same environment. We will assume the two have converged on the predictive distribution, i.e. distribution over some observables in the environment, but may have different generative models containing different latent variables. Under what conditions can one agent guarantee that their latents are a function of the other agents latents? We give simple conditions under which such translation is guaranteed to be possible: the natural latent conditions. We also show that, absent further constraints, these are the most general conditions under which translatability is guaranteed. Crucially for practical application, our theorems are robust to approximation error in the natural latent conditions.