Measuring Alignment Drift via Trajectory Prefixes

Owen Terry, Maksym Andriushchenko, “Measuring alignment drift via trajectory prefixes,” LESSWRONG 2026

Predicting how Model Propensities Generalize out of Distribution

June Hunter, Lily Wen, Owen Terry, Johannes Taraz, Ben Cohen, Niels Warncke, “Predicting how model propensities generalize out of distribution,” arXiv:2610.14937

Implications of Continual Learning for LLM Agents: Introduction

RohanS, Rauno Arike, Owen Terry, Achu Menon, Zhijing Jin, Francis Rhys Ward, Seth Herd, “Implications of Continual Learning for LLM Agents: Introduction,” LESSWRONG 2026