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

Misclassification Detection research benchmarks – Page 2 · SOTA2 Research

Misclassification Detection

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Tiny ImageNet (test)

86.1AUROC

10

Feb 26, 2026

CUPID Epistemic

85.5AUC

9

Mar 12, 2026

CUPID Aleatoric

87AUC

9

Mar 12, 2026

CIFAR-100-C 1.0 (test)

84.05AUROC

9

Feb 26, 2026

CIFAR-10-C 1.0 (test)

90.38AUROC

9

Feb 26, 2026

91.9Misclassification AUPR

8

May 26, 2026

96.17AUPR (Conf)

8

Feb 26, 2026

GoEmotions (test)

69.8AUROC

7

May 26, 2026

wnut 2017 (test)

0.911ROC-AUC

7

Apr 7, 2026

conll 2003 (test)

89.6ROC-AUC

7

Apr 7, 2026

ImageNet-1K (val)

sample difficulty-aware entropy regularization

45.69FPR@95% (MSP)

7

Feb 26, 2026

CIFAR-10-LT (ρ = 0.01)

85.99AUPR (Confidence)

6

Feb 26, 2026

0.844AUROC

3

Feb 26, 2026

AMAZON SOFTWARE (test)

0.688AUROC

3

Feb 26, 2026

AMAZON FASHION (test)

89.7AUROC (%)

3

Feb 26, 2026

88.2AUROC

2

Feb 26, 2026

Twitter POS-annotated tweets (test)

Softmax Baseline

89AUROC

2

Feb 26, 2026

WSJ (Penn Treebank) (test)

Softmax Baseline

0.96AUROC

2

Feb 26, 2026

Reuters 40 subset of 52 (test)

Softmax Baseline

91AUROC

1

Feb 26, 2026

Reuters 6 subset of Reuters 8 (test)

Softmax Baseline

89AUROC

1

Feb 26, 2026

20 Newsgroups 15 (test)

Softmax Baseline

89AUROC

1

Feb 26, 2026

Page 2 of 2

Popular datasets

CIFAR ImageNet CIFAR-100 Reuters IMDB

Follow for updates

Join SOTA2 Contact