ResearchBenchmarksMachine Unlearning Utility and Similarity Assessment on CIFAR-10Machine Unlearning Utility and Similarity Assessment on CIFAR-10Follow94.61Retained Accuracy (Acc_r)Original86.165288.357690.5592.7424Nov 24, 2025Evaluation resultsMethodMethodLinksRetained Accuracy (Acc_r)Forgetting Accuracy (Acc_f)AUS ScoreCKA FidelityCKA RetainedRUS Score (o)Original2025.1194.6194.73511001000SCRUB2025.1193.343.895319881POUR2025.1192.860.3797239585SalUn2025.1192.720.7897289281SURE2025.1192.52.695349879ProjUn2025.1186.4920.476348976