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

Hallucination Detection research benchmarks – Page 11 · SOTA2 Research

Hallucination Detection

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

21.3HaluEval Delta

10

May 20, 2026

61.11Consistency Rate

10

Mar 26, 2026

92.67Summary Consistency Rate

10

Mar 26, 2026

53.5Object-Relation Accuracy

10

Apr 15, 2026

HotpotQA-WC (test)

0.86AuROC

10

Feb 26, 2026

100AuROC

10

Feb 26, 2026

Winobias (test)

54.43AUROC

10

Feb 26, 2026

Winogrande (test)

52.48AuROC

10

Feb 26, 2026

94.15AuROC

10

Feb 26, 2026

0.546AuROC

10

Feb 26, 2026

Generated responses (test)

Fisher-based label propagation

92.5Accuracy

10

Feb 26, 2026

Glaive General Mixed

L26 Smoothness (single)

98.2Recall

10

Feb 26, 2026

RAGTruth Llama-13B

89.47Recall

10

Feb 26, 2026

RAGTruth Llama-7B

92.54Recall

10

Feb 26, 2026

90F1-R

10

Feb 26, 2026

Unified Hallucination Detection (test)

0.812AUROC

9

Jun 9, 2026

Non-Contradiction

0.587AUROC

9

Jun 2, 2026

TriviaQA (out-of-distribution)

80.44AUC

9

Apr 21, 2026

Hallucination Detection inference time comparison

2.511Feature Extraction Time (s)

9

Apr 6, 2026

66.7AUROC

9

Mar 26, 2026

60.8AUROC

9

Mar 26, 2026

FAVA-Annotations (test)

LLM-Check Attn Score

68.19AUROC

9

Feb 26, 2026

ActivityNet FOIL

93.5Accuracy

9

Feb 26, 2026

GASP+base (trained)

67.9AUC

8

Jul 7, 2026

0.506PRR

8

Jun 18, 2026

Page 11 of 17

Popular datasets

RAGTruth HaluEval TriviaQA SQuAD CoQA

Follow for updates

Join SOTA2 Contact