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

CRPE

Benchmarks

Task NameDataset NameSOTA ResultTrend
Hallucination EvaluationCRPE relation
Accuracy79.1
23
Hallucination EvaluationCRPE
Score75.6
14
Multimodal Hallucination EvaluationCRPE
Existence Score96.8
13
Compositional Reasoning and Perception EvaluationCRPE
Exist Score94.15
13
Spatial ReasoningCRPE
Subject Accuracy82.2
12
Spatial UnderstandingCRPE
Accuracy77.15
11
Discriminative Hallucination DetectionCRPE R
Accuracy70.7
10
Vision-Centric UnderstandingCRPE
Accuracy77
9
Spatial ReasoningCRPE
Accuracy80.21
7
3D TaskCRPE
Accuracy80.21
7
General Visual Question AnsweringCRPE relation
Score79.2
5
Relation ComprehensionCRPE (Subject, Predicate, Object)
Subject Accuracy69.21
3
Object RecognitionCRPE Existence
Existence Accuracy92.14
3
Showing 13 of 13 rows