Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

MLDR

Benchmarks

Task NameDataset NameSOTA ResultTrend
Long-context RetrievalMLDR
nDCG@1062.86
36
RetrievalMLDR (test)
NDCG@1035
34
Fragment RetrievalMLDR (val)
Precision84.02
29
Multilingual Long-context RetrievalMLDR
nDCG@1074.4
28
Multilingual Long Document RetrievalMLDR 13 (test)
NDCG@1057.5
18
multilingual long-doc retrievalMLDR (test)
Average Retrieval Score65
14
Long-context retrievalMLDR en
nDCG@100.783
11
Information RetrievalMLDR
Recall@10100
10
Document RetrievalMLDR
nDCG@10 (German)28.6
10
Document RerankingMLDR zh
P@195.1
9
RetrievalMLDR European Languages
nDCG@1075.8
8
RetrievalMLDR All Languages
nDCG@1068.6
8
Long-form Dialogue DecompositionMLDR (test)
Pairwise Precision81.2
3
Showing 13 of 13 rows