TY  - RPRT
TI  - A Monosemantic Attribution Framework for Stable Interpretability in Clinical Neuroscience Transformer-Based Language Models
AU  - Michail Mamalakis
AU  - Tiago Azevedo
AU  - Cristian Cosentino
AU  - Chiara D'Ercoli
AU  - Subati Abulikemu
AU  - Zhongtian Sun
AU  - Richard Bethlehem
AU  - Pietro Lio
PY  - 2026
UR  - https://arxiv.org/abs/2601.17952
ID  - 2601.17952
ER  -