Machine Heart

Apr 22, 2026 · Artificial Intelligence

ProSafePrune: One‑Shot Pruning to Eliminate Over‑Refusal in Large Language Models

ProSafePrune, a low‑rank parameter pruning framework presented at ICLR 2026, precisely removes over‑harmful encoding in LLMs, dramatically reducing over‑refusal while preserving safety defenses and slightly improving general‑task performance.