51 citations · 120 across the 10 of their papers we have counts for
1 paper · 1 filter
Md Sazid Uddin, Md. Khairul Alam Mazumder, M. F. Mridha
Mechanistic edits (ablations, weight edits, activation steering) are the standard tools for unlearning a harmful capability from a neural network while preserving useful ones. Curr…