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

DB

Benchmarks

Task NameDataset NameSOTA ResultTrend
Code Correctness EvaluationDB
F1 Score68.6
24
Database Querying Agent TaskDB
Success Rate70.1
19
InstanceOf triple classificationDB99K-242 (test)
Accuracy94.21
13
Private Information RetrievalDB 4GB (test)
QPS2,201
9
Private Information RetrievalDB 1GB (test)
QPS8,244
9
Link predictionDB100k
MRR0.419
9
Private Information RetrievalDB 2GB (test)
Throughput (QPS)723.5
4
Topic ClassificationDB
Accuracy98.5
4
Showing 8 of 8 rows