NQ
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
NQ
50.979NDCG@5
14
NQ
56.3T Metric
14
NQ
59.46EM
14
NQ-Open
0.277ROUGE-Lsum
14
NQ
75.8Recall
13
NQ
47.6EM
13
NQ
43EM
13
NQ
56.1EM
12
NQ
100Recall
12
NQ
57.8Robustness RACC (Paradox)
12
NQ
0.945Cohen's d
12
NQ
0.636Spearman's rho
12
NQ
5,224Throughput (Samples/sec)
12
NQ
71.2CEM
12
NQ
89.85Recall@10
12
NQ Dataset
45.36EM (Long Question)
12
NQ in-corpus setting (test)
39Accuracy
12
NQ BEIR (test)
0.69nDCG@10
12
NQ entity-swapped (test)
73.73Exact Match
12
NQ
52.4EM
11
NQ
62Accuracy
11
NQ new documents
69.3R@1
11
NQ initial documents
69.6R@1
11
NQ (test)
77.01Top-20 Accuracy
11
NQ (test)
44.11r_connect
10