MIST at OdenseNLP

Assistant Professor Lukas Galke Poech from OdenseNLP is leading MIST: Scalable Mechanistic Interpretability for Safe and Trustworthy LLM Agents, a project focused on making language model agents better understood, safer, and more reliable in practice.

MIST develops scalable methods for mechanistic interpretability and studies the causal mechanisms behind key agent capabilities such as instruction following, tool use, planning, reasoning, and multi-agent communication. The project also investigates how safety interventions propagate through multi-agent systems, with the long-term goal of supporting robust evaluation and safety certification for LLM agents.

Read more at the MIST project page. MIST is funded by the Novo Nordisk Foundation through a RECRUIT grant (2026-2031).