Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Machine Unlearning research benchmarks – Page 15 · SOTA2 Research

Machine Unlearning

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

64.62Accuracy (Bio)

8

May 19, 2026

99.8VerbMem

8

May 14, 2026

58.02Overall Score

8

Apr 28, 2026

CIFAR-10 class-wise

99.71Accuracy Difference (Df)

8

Apr 15, 2026

CIFAR-100 10% Random Forgetting VGG16-BN

99.51Retention Accuracy (RA)

8

Apr 14, 2026

CIFAR-100 30% random forgetting

99.9Utility (D_r^train)

8

Apr 13, 2026

TinyImageNet 30% random data forgetting

68.7Dr (Train Set)

8

Apr 13, 2026

0.522Unlearn Score

8

Mar 31, 2026

8.28Relevance Score

8

Mar 12, 2026

TOFU (Forget05)

6.62Relevance

8

May 27, 2026

0.029MIA

8

Feb 26, 2026

87.4USR

8

Feb 26, 2026

MUSE-Books Harry Potter

Mistral-7B (base)

48.4R-forget

8

Feb 26, 2026

ImageNet100 30% influence-style deletion (retain)

71.8Retention Accuracy

8

Feb 26, 2026

CIFAR-10 Influence-inspired deletion (test)

67.91Accuracy

8

Feb 26, 2026

CIFAR-10 High-gradient-norm deletion (test)

68.25Accuracy

8

Feb 26, 2026

CIFAR-10 Low-margin deletion (test)

68.12Accuracy

8

Feb 26, 2026

CIFAR-10 High-loss deletion (test)

68.18Acc

8

Feb 26, 2026

MUSE-Bench raw samples (test)

32.13Knowledge Forgetting

8

Feb 26, 2026

Harry Potter Extended (test)

39.99Knowledge Retention (f)

8

Feb 26, 2026

Internal e-commerce benchmark small-scale seller, 66 items

86.4Forget Set ROUGE

8

Feb 26, 2026

internal e-commerce benchmark large-scale seller 1065 items (Forget Set)

48.9ROUGE

8

Feb 26, 2026

7.61FID

8

Feb 26, 2026

MLLMU-Bench (retain set)

46.35Accuracy

7

Jul 9, 2026

0.3Accuracy

7

Jun 18, 2026

Page 15 of 22

Popular datasets

CIFAR TOFU ImageNet WMDP MUSE

Follow for updates

Join SOTA2 Contact