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

Trivia

Benchmarks

Task NameDataset NameSOTA ResultTrend
Single-Hop Question AnsweringTrivia
Exact Match68.1
30
Selective Question AnsweringTRIVIA
AUC-F1 (Coverage)92.5
12
Factoid Question AnsweringTrivia Estonian
Accuracy42.9
10
Retrieval Question AnsweringTrivia
MRR47.6
6
Retrieval Question AnsweringTrivia (in-domain)
MRR0.555
4
Showing 5 of 5 rows