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

BEIR

Benchmarks

Task NameDataset NameSOTA ResultTrend
Information RetrievalBEIR
SciFact0.9954
174
Information RetrievalBEIR (test)
TREC-COVID Score91.6
130
Information RetrievalBEIR v1.0.0 (test)
ArguAna72.9
75
Re-rankingBEIR (test)
NQ59.8
46
RankingBEIR selected subset v1.0.0 (test)
TREC-COVID84.7
38
Zero-shot Information RetrievalBEIR
NFCorpus NDCG@10 (Zero-shot)39.1
38
RerankingBEIR
NQ NDCG@50.5106
35
Text RankingBEIR out-of-domain
Average Score (BEIR OOD)56.2
27
Information RetrievalBEIR COVID v1 (test)
nDCG@1063.5
26
Document RerankingBEIR
NDCG@10 (Covid)85.51
24
Document RerankingBEIR (test)
NDCG@10 (NFCorpus)0.378
23
Information RetrievalBEIR v1 (test)
ArguAna49.2
22
Information RetrievalBEIR-7 Average
NDCG@10 (Dense)67.1
21
Document RetrievalBEIR 15
NDCG@100.443
21
OOD document detectionBEIR (test)
ArguAna1
20
Information RetrievalBEIR v1.0 (test)
FEVER Score94.7
20
RerankingBEIR (test)
Covid Score78.5
19
Passage RerankingBEIR
nDCG@10 (COVID)85.77
18
Document RerankingBEIR
Average NDCG@1055.4
17
Relevance RankingBEIR 9-dataset (test)
NDCG@10 (NFCorpus)42.33
16
Semantic RelatednessBEIR Semantic Relatedness Tasks (test)
ArguAna Score47.9
16
Re-rankingBEIR OOD
Ar Score41
15
RerankingBEIR
NDCG@10 (Avg)65.11
14
Text RetrievalBEIR-5 all-MiniLM-L6-v2 (test)
Average NDCG@1052.3
14
Information Retrieval RerankingBeIR collection 12 datasets
nDCG@1060.2
13
Showing 25 of 64 rows