Papers
Our published papers and preprints, newest first, each with a line in plain English. All are free to read.
-
arXiv preprintSeptember 2026
Robot World Models Are Not Invariant to How the Actions Are Written
A robot world model trained on one way of writing actions breaks when it’s handed the same actions written another, equivalent way.
World models and sciencearXiv:2609.23252
-
arXiv preprintSeptember 2026
Exact Finite Attention Responses From RoPE Derivatives
An exact formula for how a model’s attention responds when parts of its input are moved, removed or changed.
In-context learningarXiv:2609.14127
-
ICML 2026June 2026
Predictable Compression Failures: Order Sensitivity and Information Budgeting for Evidence-Grounded Binary Adjudication
Treats hallucination as a measurable shortfall of information, and gives a rule for when a model should answer or abstain.
Hallucination and verificationarXiv:2509.11208
-
arXiv preprint2025, revised 2026
LLMs are Bayesian in Expectation, Not Realization
Averaged over the order of their examples, language models come close to ideal Bayesian reasoning; any single ordering can drift from it.
In-context learningarXiv:2507.11768
-
Open textbook2026
Information Geometry for Generative Models
A free 13-chapter textbook on the maths behind generative AI, from compression and Bayesian prediction to transformers and diffusion.
github.com/leochlon/teta
Talk slides are on the About page. For a paper we haven’t listed, or a PDF that won’t open, email us.