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

EventHallusion

Benchmarks

Task NameDataset NameSOTA ResultTrend
Hallucination DetectionEventHallusion
Accuracy75
30
Video Hallucination EvaluationEventHallusion
Entire Score84.2
29
Free-form descriptionEventHallusion
LLM Likert Score2.52
24
Hallucination DetectionEventHallusion Entire, Mix, Misleading
Accuracy (Entire)69.3
22
Hallucination-oriented Video UnderstandingEventHallusion (test)
Accuracy91.9
22
Video UnderstandingEventHallusion
Accuracy73.84
18
Hallucination ExaminationVidHalluc, VideoHallucer, EventHallusion
VidHalluc Score81.2
17
Hallucination examinationEventHallusion
Average Score69.19
15
Hallucination EvaluationEventHallusion binary QA (test)
Accuracy0.655
15
Detailed Description MatchingEventHallusion
Accuracy (Entire)54.6
13
Video Question AnsweringEventHallusion
Accuracy (Entire)70.2
13
Hallucination EvaluationEventHallusion 36
Entire Score62.28
5
Showing 12 of 12 rows