Safety Edits That Don't Stick: datasets Collection ValueShift Research, Lexsi hackathon: RMU unlearning and Antidote repair on Tiny Aya. Repair artifacts are deliberately harmful. • 4 items • Updated 3 days ago
Safety Edits That Don't Stick: models Collection ValueShift Research, Lexsi hackathon: RMU unlearning and Antidote repair on Tiny Aya. Repair artifacts are deliberately harmful. • 3 items • Updated 3 days ago
crosslingual-rule-following/crv-norule-baseline-responses-all Viewer • Updated 22 days ago • 2.6k • 159
crosslingual-rule-following/crv-norule-baseline-responses-all Viewer • Updated 22 days ago • 2.6k • 159