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

AQUA-RAT

Benchmarks

Task NameDataset NameSOTA ResultTrend
Mathematical ReasoningAQUA-RAT
Accuracy91.73
153
Mathematical ReasoningAQuA-RAT (test)
Accuracy83
40
Mathematical CalculationAQuA-RAT
Accuracy (AQuA-RAT)83.1
20
Algebraic ReasoningAQuA-RAT
Accuracy66.8
16
Algebraic Question AnsweringAQUA-RAT Synthetic NIID 1.0 (test)
Accuracy28
7
Algebraic Question AnsweringAQUA-RAT Synthetic IID 1.0 (test)
Accuracy29.9
7
Mathematical ReasoningAQuA-RAT Multiple Client (test)
Client 1 Accuracy64.92
6
Mathematical ReasoningAQuA-RAT Single Client
Accuracy0.6102
6
Math reasoningAQuA-RAT
Accuracy59.06
3
Mathematical ReasoningAQUA-RAT STREET
Answer Accuracy78
3
Mathematical ReasoningAQUA-RAT standard (test)
ACC (%)40.16
3
Algebraic ReasoningAQuA-RAT
PTR Advance Rate100
2
Showing 12 of 12 rows