Representation-Aware Unlearning via Activation Signatures: From Suppression to Entity-Signature Erasure 文章

ArXiv CS.CL2026-05-27NEWSen作者: Syed Naveed Mahmood, Md. Rezaur Rahman Bhuiyan, Tasfia Zaman, Jareen Tasneem Khondaker, Md. Sameer Sakib, K. M. Shadman Wadith, Nazia Tasnim, Farig Sadeque

详细信息

来源站点
ArXiv CS.CL
作者
Syed Naveed Mahmood, Md. Rezaur Rahman Bhuiyan, Tasfia Zaman, Jareen Tasneem Khondaker, Md. Sameer Sakib, K. M. Shadman Wadith, Nazia Tasnim, Farig Sadeque
文章类型
NEWS
语言
en
发布日期
2026-05-27

摘要

arXiv:2601.10566v5 Announce Type: replace Abstract: Entity-level unlearning is usually evaluated by what a model says: whether it stops naming the target, refuses a query, or shifts a Truth Ratio distribution. These output-level tests, however, do not show whether a subject's internal representation has been attenuated. We introduce the Entity Representation Unlearning Framework (ERUF), a representation-aware framework that mines subject-specific activation signatures, suppresses the corresponding activation direction, and distills the behavior into LoRA parameters. Among evaluated baselines, ERUF is the only method that jointly achieves surface-level suppression, internal attenuation, and utility preservation. On TOFU forget10, ERUF achieves FQ = 0.99 and MU = 0.62, matching reported oracle utility while approaching oracle forget quality. Across most standard foundation-model settings, ERUF maintains low leakage and low internal target activation, with SMR between 0.00% and 1.