Towards Isolated Interventions via Almost Orthogonal Features in Language Models

Moritz Miller and two co-authors submitted the paper "Towards Isolated Interventions via Almost Orthogonal Features in Language Models" to arXiv on February 4, 2026, with a revised version posted on July 9, 2026. The paper is categorized under cs.LG and explores methods for achieving isolated interventions in language models using approximately orthogonal features.