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

XQuAD

Benchmarks

Task NameDataset NameSOTA ResultTrend
Multilingual Information RetrievalXQuAD
Completion@1078.82
80
Label ProjectionXQuAD
COMET Score87.3
48
Question AnsweringXQuAD
F1 (de)80.5
21
Multilingual UnderstandingXQuAD (test)
Accuracy49.47
12
Question AnsweringXQuAD
Accuracy64.26
12
Question AnsweringXQuAD 1.0 (test)
F1 Score79.6
10
Performance PredictionXQUAD
MAE3.15
9
Zero-shot performance predictionXQUAD
MAE2.89
9
Question AnsweringXQuAD (test)
F1 Score82.4
9
Question AnsweringXQuAD English Distractor I - (X), Q(X & EN) (Avg)
Language Consistency98.23
8
Question AnsweringXQuAD Code Switched P, I - (EN), Q(X) (Avg)
Language Consistency100
8
Question AnsweringXQuAD Monolingual P, I, Q - (X) (Avg)
Language Consistency100
8
Question AnsweringXQuAD v1.1 (test)
F1 (en)89
8
Question AnsweringXQuAD seen languages XTREME (test)
F1 Score74.7
6
Question AnsweringXQuAD
English QA Score83.9
6
Question AnsweringXQuAD vi
Score42.2
4
Question AnsweringXQuAD zh
Raw Score27.24
4
Multilingual Question Answeringxquad vi
Normalized Performance89.29
3
Multilingual Question Answeringxquad zh
Normalized Performance60.28
3
Question AnsweringXQuAD languages not in MLQA
F1 (el)81.9
3
Showing 20 of 20 rows