English

Kinetic theory for Transformers and the lost-in-the-middle phenomenon

Analysis of PDEs 2026-05-12 v1 Machine Learning Probability

Abstract

We study causal self-attention dynamics -- a toy model for decoder Transformers -- which we interpret as a non-exchangeable interacting particle system. Adapting cumulant expansions to the triangular causal dependency structure of the model, and appealing to non-hierarchical methods to estimate correlations using Glauber calculus, we prove a quantitative mean-field limit result and a next-order characterization of correlations. For iid uniformly distributed tokens, the limiting correlation equation can be solved in closed form and we obtain a rigorous explanation of the empirically observed \emph{lost-in-the-middle} phenomenon: the token retrieval profile, as a function of the source position in the prompt, is U\mathsf{U}-shaped, with primacy, recency, and a unique interior minimum under an explicit smallness condition.

Keywords

Cite

@article{arxiv.2605.09213,
  title  = {Kinetic theory for Transformers and the lost-in-the-middle phenomenon},
  author = {Mitia Duerinckx and Borjan Geshkovski and Stefano Rossi},
  journal= {arXiv preprint arXiv:2605.09213},
  year   = {2026}
}