2 papers
cs.LG2026
Uncovering Model Processing Strategies with Non-Negative Per-Example Fisher Factorization
Michael Matena, Colin Raffel
We introduce NPEFF (Non-Negative Per-Example Fisher Factorization), an interpretability method that aims to uncover strategies used by a model to generate its predictions. NPEFF de…
cs.CL2026
Better and Worse with Scale: How Contextual Entrainment Diverges with Model Size
Dikshant Kukreja, Kshitij Sah, Gautam Gupta +5
Larger language models become simultaneously better and worse at handling contextual information -- better at ignoring false claims, worse at ignoring irrelevant tokens. We formali…