CAO, Zouying; YANG, Yifei; ZHAO, Hai. SCANS: Mitigating the Exaggerated Safety for LLMs via Safety-Conscious Activation Steering. Proceedings of the AAAI Conference on Artificial Intelligence, [S. l.], v. 39, n. 22, p. 23523–23531, 2025. DOI: 10.1609/aaai.v39i22.34521. Disponível em: https://ojs.aaai.org/index.php/AAAI/article/view/34521. Acesso em: 10 may. 2026.