SQuAD
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
SQUAD Open (test)
56.6Exact Match
39
SQuAD v2
93F1 Score
38
SQuAD KRE-curated version
72.6F1 Score
36
SQuAD v2
1ASR Score
36
SQuAD
71.4F1 Score
36
SQuAD
2,075Correlation (alpha=0.17)
35
SQuAD
89.3Score
35
SQuAD 2.0
92.9F1 Score
34
SQuAD 2.0 (test)
89.7EM
34
SQuAD
83AUROC
32
SQuAD
99.6Accuracy
32
SQuAD
5.87ECE
31
SQuAD
25.8Exact Match (EM)
30
SQuAD
99AUROC
30
SQuAD v2.0
57.82Disagreement
30
SQuAD Open-domain 1.1 (test)
61.8Exact Match (EM)
30
SQuAD 1.1 (test)
25.8BLEU-4
29
SQuAD v2
0.798ROC-AUC
28
SQuAD
69.9AUCs
28
SQuAD-Open
59.5EM
28
SQuAD v2
0ASR
27
SQUAD
86.3Accuracy (ACC)
27
SQuAD v2
1,853Correlation
26
SQuADShifts (test)
0.883AUC
26
SQuAD (val)
91.67F1 Score
26