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

MLLM Unlearning research benchmarks · SOTA2 Research

MLLM Unlearning

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

MLLMU-Bench Retain Set 10% ratio

34.4Cloze Accuracy

30

Jun 2, 2026

MLLMU-Bench 10% ratio (test)

40Cloze Accuracy

30

Jun 2, 2026

MLLMU-Bench forget set, 10% ratio

40Cloze Accuracy

30

Jun 2, 2026

MLLMU 15% Task (Real)

33.2Rouge

5

Feb 26, 2026

MLLMU 15% Task (retain)

52.7ROUGE

5

Feb 26, 2026

MLLMU 15% Task (Generalization)

37.2Accuracy

5

Feb 26, 2026

MLLMU 15% Task (forget)

34.7Accuracy

5

Feb 26, 2026

MLLMU 10% Task (Real)

24.6Rouge

5

Feb 26, 2026

MLLMU 10% Task (retain)

49.7ROUGE

5

Feb 26, 2026

MLLMU 10% Task (Generalization)

36.5Accuracy

5

Feb 26, 2026

MLLMU 10% Task (Forget)

36.9Accuracy

5

Feb 26, 2026

MLLM Unlearning

—

—Primary metric

0

Feb 26, 2026

Popular datasets

MLLMU MLLMU-Bench MLLM Unlearning

Follow for updates

Join SOTA2 Contact