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

Fact*

Benchmarks

Task NameDataset NameSOTA ResultTrend
Fact VerificationFACT
Accuracy99.44
15
Answerability PredictionFACT n=10 (matched pairs)
AUC0.75
9
Hallucination DetectionFact*
AUC-ROC78.6
4
Graph completionFact
Chain Precision100
3
Showing 4 of 4 rows