Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Hallusion

Benchmarks

Task NameDataset NameSOTA ResultTrend
Multimodal ReasoningHallusion
Score63.51
13
Visual PerceptionHallusion
Accuracy72.45
10
Hallucination MitigationHallusion
Accuracy64.56
10
Multimodal ComprehensionHallusion
Score50.9
8
Visual Hallucination DetectionHallusion-VD
Accuracy63.3
4
Showing 5 of 5 rows