TY  - RPRT
TI  - Leak@$k$: Unlearning Does Not Make LLMs Forget Under Probabilistic Decoding
AU  - Hadi Reisizadeh
AU  - Jiajun Ruan
AU  - Yiwei Chen
AU  - Soumyadeep Pal
AU  - Sijia Liu
AU  - Mingyi Hong
PY  - 2026
UR  - https://arxiv.org/abs/2511.04934
ID  - 2511.04934
ER  -