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

RePOPE

Benchmarks

Task NameDataset NameSOTA ResultTrend
Hallucination AttributionRePOPE
Insertion Score80.55
16
Hallucination InterpretationRePOPE
Insertion91.14
16
Discriminative Hallucination DetectionRePOPE
Accuracy93.8
10
Visual Attribution-based Hallucination RepairRePOPE
Repair Rate96
8
Showing 4 of 4 rows