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

ESCI

Benchmarks

Task NameDataset NameSOTA ResultTrend
Adversarial Attack Effectiveness in RAG PipelineESCI position 6 (test)
Success Rate @10100
28
GEO Attack DetectionESCI Pipeline (≈1:9) (test)
FDR0
28
GEO Attack DetectionESCI Balanced (1:1) (test)
FDR0
28
Search RelevanceESCI
Macro F161.03
14
Relevance ClassificationESCI-ES (test)
ACC68.1
13
Relevance ClassificationESCI-US (test)
Accuracy70.23
13
Product SearchESCI
NDCG (Two-C)82.59
9
Product SearchESCI
NDCG@1058.8
8
Query RewritingESCI
Coverage (cov(Q))171
6
Product SearchESCI (test)
NDCG (Two-C)-
0
Showing 10 of 10 rows