English

Exploring Geographic Relative Space in Large Language Models through Activation Patching

Machine Learning 2026-05-15 v1

Abstract

The increased use of Large Language Models (LLMs) in geography raises substantial questions about the safety of integrating these tools across a wide range of processes and analyses, given our very limited understanding of their inner workings. In this extended abstract, we examine how LLMs process relative geographic space using activation patching, an emerging tool for mechanistic interpretability.

Keywords

Cite

@article{arxiv.2605.14535,
  title  = {Exploring Geographic Relative Space in Large Language Models through Activation Patching},
  author = {Stef De Sabbata and Rahul Baiju and Stefano Mizzaro and Kevin Roitero},
  journal= {arXiv preprint arXiv:2605.14535},
  year   = {2026}
}