Belebele
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Belebele FLORES (test)
0.98Pearson Correlation
81
Belebele
0.932Comp@10
80
Belebele
0.772Mean AUROC
48
Belebele
0.926nDCG@20
45
Belebele
84.7Accuracy
39
Belebele 28 European languages
85.91Overall Score
34
Belebele c
37.11Accuracy (Normalized)
32
BELEBELE
80Average RC Score (BELEBELE)
31
BelebeleRetrieval
96.26nDCG@10
26
Belebele Target language
55.78MRC Score
24
Belebele EN
75.33Accuracy
22
BELEBELE 5 Unseen Languages
85BELE Score
20
BELEBELE 8 Seen Languages
81.8BELE Score
20
Belebele
79.8Accuracy
18
Belebele English
94.56Accuracy
18
Belebele Source language en
89.8MRC Score
16
Belebele Ukrainian (test)
89.56Accuracy
14
Belebele Spanish (test)
91Accuracy
14
Belebele Polish
87.56Accuracy
13
BELEBELE Indonesian
67.1Accuracy (Target Language)
13
Belebele
81.9nDCG@10 (Afr)
12
Belebele
92.7Afr Score
12
Belebele Hindi
84Accuracy
12
Belebele Korean (test)
90Accuracy
12
Belebele French
74.2Score
12