BRIGHT
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
BRIGHT
54.8Mean nDCG@10
94
BRIGHT
40.3NDCG@10 (Avg)
54
BRIGHT 1.0 (test)
37.9nDCG@10 (Avg)
41
BRIGHT
33.9BRIGHT Score (Biology)
32
BRIGHT StackExchange
62.3Biology Score
29
BRIGHT
20Biology nDCG@5
24
BRIGHT v1 (test)
49.1nDCG@10 (Avg)
22
BRIGHT (test)
50.1Bio Score
20
BRIGHT 12 datasets aggregate (test)
12.74NDCG@10
20
BRIGHT 1.0 (test)
61.77NDCG@10 (Bio.)
16
BRIGHT (test)
72.2Accuracy (Biology)
15
BRIGHT (test)
54.5nDCG@10 (Biology)
13
BRIGHT static retrieval setting PRO
33.8NDCG@10 (Overall)
13
BRIGHT
65.8nDCG@1 (Econ)
13
BRIGHT v1 (leaderboard)
46.8Average Retrieval Score
12
BRIGHT 6class (test)
43.4Accuracy
11
BRIGHT
91.71F1 (bcd)
10
BRIGHT BEIR
60NDCG@10 (Bio)
8
BRIGHT unseen 6 subsets (test)
11.79nDCG@10
7
BRIGHT DFC25-T2
43.75mIoU
6
BRIGHT SO
23.3Recall@10
3
BRIGHT Econ
19.1Recall@10
3
BRIGHT ES
39.2Recall@10
3
BRIGHT Earth Science
0.722nDCG@10
3
BRIGHT (test)
0.646Accuracy
3