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 Mitigation research benchmarks · SOTA2 Research

Hallucination Mitigation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Retrieval-Augmented Logical Reasoning

95.2Accuracy

31

Feb 26, 2026

75CHAIR_S

24

Feb 26, 2026

HallusionBench visual-dependent setting (test)

33.21qAcc (Question Accuracy)

21

May 14, 2026

22.1HSR

15

Feb 26, 2026

Factual Grounding and Causal Reasoning Evaluation Set

4.25AC

14

Feb 26, 2026

92.5Faithfulness Score

12

Jun 16, 2026

98.4Faithfulness

12

Jun 16, 2026

99Faith

12

Jun 16, 2026

52.1F1 Score

12

Feb 26, 2026

MME Existence, Count, Position, Color

195Existence Score

12

Feb 26, 2026

88.2ACC

10

Feb 26, 2026

64.56Accuracy

10

Feb 26, 2026

5.36SPI

9

Apr 14, 2026

Delulu N=1,950, 7 languages (held-out)

Qwen2.5-Coder-7B

61.8Exact Match (EM)

8

Jun 4, 2026

Lyrics Scenario Aggregated

0.081Hallucination Rate

8

Apr 21, 2026

Code-switch Scenario Aggregated

0.261Hallucination rate

8

Apr 21, 2026

English Scenario Aggregated

Qwen3-Omni-Inst

0.7Hallucination Rate

8

Apr 21, 2026

Dialect Scenario Aggregated

11.7Hallucination Rate

8

Apr 21, 2026

Mandarin Scenario Aggregated

0.2Hallucination Rate

8

Apr 21, 2026

95.98HR (%)

6

Jun 8, 2026

32.7Hallucination Rate

6

Apr 7, 2026

26.78CHAIRi Score

4

Mar 12, 2026

Popular datasets

RAGTruth SHR UrbanSound CHAIR POPE

Follow for updates

Join SOTA2 Contact