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 Evaluation research benchmarks – Page 4 · SOTA2 Research

Hallucination Evaluation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

LaMini-Hallucination 1.0 (test)

40Total

10

Feb 26, 2026

Style-POPE Overall Average

79.1F1 Score

9

Jul 7, 2026

Style-POPE Random

0.844F1 Score

9

Jul 7, 2026

Style-POPE Adversarial

75.3F1 Score

9

Jul 7, 2026

64Accuracy

9

Jul 7, 2026

Thermal-RGB images

85Accuracy

9

Jul 7, 2026

79Accuracy

9

Jul 7, 2026

82Accuracy

9

Jul 7, 2026

61Accuracy

9

Jul 7, 2026

2.57Hallucination Score

9

Apr 28, 2026

Style-CHAIR Original

8.8CHAIRi Score

8

Jul 7, 2026

Style-CHAIR Sketch

11.4CHAIRi Score

8

Jul 7, 2026

Style-CHAIR Painting

11.2CHAIRi Score

8

Jul 7, 2026

Style-CHAIR Graffiti

12CHAIRi Score

8

Jul 7, 2026

Hallucination Evaluation Suite (Object HalBench, AMBER, HallusionBench)

5.84Avg. Hallucination Score

8

Jun 30, 2026

CHAIR instance-level (test)

11.7Brightness

8

Mar 30, 2026

71.1Accuracy

8

Feb 26, 2026

6.87Correctness

8

Feb 26, 2026

curated hallucination benchmark

PP-OCRv6_medium

93.2Accuracy

7

Jun 12, 2026

Hallucinations evaluation 1.0 (test)

52.38AI2D Score

7

Mar 31, 2026

SPD-Faith Bench

43.8CR

7

Feb 26, 2026

58.1Hallucination Score

7

Feb 26, 2026

9.33Relevancy

6

May 7, 2026

AMBER AttackVLM (adversarially perturbed)

24.4CHAIR

6

Apr 23, 2026

AMBER Generative Subset benign setting

Qwen-VL-Chat + ORCA

3.2CHAIR Score

6

Apr 23, 2026

Page 4 of 6

Popular datasets

POPE MME CHAIR Style-CHAIR HallusionBench

Follow for updates

Join SOTA2 Contact