一作ICLR 2026Poster
No, of Course I Can! Deeper Fine-Tuning Attacks That Bypass Token-Level Safety Mechanisms
Joshua Kazdan · Abhay Puri · Rylan Schaeffer · Lisa Yu · Chris Cundy · Jason Stanley · Sanmi Koyejo · Krishnamurthy Dvijotham
首次获取该候选人的完整档案及 StartLux 评估共 20 积分;每位用户单独解锁,之后查看免费。主动刷新 20 积分。
实时查找电子邮箱,费用不超过20积分;未找到邮箱不扣费。
Joshua Kazdan · Abhay Puri · Rylan Schaeffer · Lisa Yu · Chris Cundy · Jason Stanley · Sanmi Koyejo · Krishnamurthy Dvijotham