作者位次 #4/4NeurIPS 2026Poster
AnchorRep: Defending LLMs Against Cross-Model Adversarial Transfer via Representation Repulsion
Gal Wertheizer · Rom Himelstein · Tomer Peretz · Avi Mendelson
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。
Gal Wertheizer · Rom Himelstein · Tomer Peretz · Avi Mendelson
Liran Cohen · Yaniv Nemcovsky · Avi Mendelson
Understanding how large language models (LLMs) store, retain, and remove knowledge is critical for interpretability, reliability, and privacy compliance. We reveal a key phenomenon: machine unlearning imprints distinct geometric signatures in the model’s input loss landscape (ILL), with unlearned...
Rom Himelstein · Amit LeVi · Brit Youngmann · Yaniv Nemcovsky · Avi Mendelson