Natural Questions
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
EVOUNA-NaturalQuestions
97.9F1 Score
35
Natural Questions (NQ) In-Domain
85.16AUROC
33
Natural Questions (NQ) (test)
47.5EM
33
NQ (Natural Questions)
51.4EM
33
Natural Questions (NQ)
45.68Exact Match (EM)
32
NQ (Natural Questions)
46.1EM
32
Natural Questions
52.2F1 Score
31
Natural Questions
92.53AR@5
31
Natural Questions
50.52Qwen2.5-3B-Instruct F1
31
Natural Questions (test)
92.53AR@5
31
NQ (Natural Questions) in-domain (val/test)
50.9Exact Match
30
NQ (Natural Questions)
42.5EM
28
Natural Questions (NQ)
66.59Top-10 Accuracy
28
Natural Questions (test)
29.9Accuracy
27
Natural Questions (NQ)
83NQ Score
26
Natural Questions (test)
2.916Speedup Ratio
26
Natural Questions
30.6Accuracy
26
Natural Questions (NQ)
36.71EM
24
Natural Questions (NQ)
65.2HSPP
23
Natural Questions (NQ)
48.96Accuracy
23
Natural Questions MRQA
81F1 Score
22
Natural Questions (NQ)
40.8EM
21
NQ (Natural Questions)
49.2Exact Match (EM)
20
Natural Questions (NQ) (formal)
79Percentage of AmE Generations
20
NQ (Natural Questions)
54.06F1 Score
20