AQUA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
AQuA
83.67Source Library Accuracy
5
AQUA
94.4Answer Selection Rate (ASR)
4
AQuA-50 (pilot)
88Accuracy
3
AQuA
93.42Mean Accuracy
3
AQUA
22.7PPL
3
AQuA
90CSR
3
AQuA
21.34PPL
3
AQUA
33.5Accuracy (format-specific prompt)
2