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

Arithmetic Addition

Benchmarks

Task NameDataset NameSOTA ResultTrend
Arithmetic AdditionArithmetic Addition Base 5, OOD-10x
Token Accuracy17.2
4
Arithmetic AdditionArithmetic Addition Base 5, OOD-5x
Token-wise Accuracy17.3
4
Arithmetic AdditionArithmetic Addition Base 5, OOD-2x
Token-wise Accuracy18.8
4
Arithmetic AdditionArithmetic Addition Base 5, ID n=5
Token-wise Accuracy23
4
Arithmetic AdditionArithmetic Addition Base 3 OOD-10x
Token-wise Accuracy31
4
Arithmetic AdditionArithmetic Addition Base 3 OOD-5x
Token-wise Accuracy30.7
4
Arithmetic AdditionArithmetic Addition Base 3 OOD-2x
Token-wise Accuracy31.4
4
Arithmetic AdditionArithmetic Addition Base 3 ID n=5
Token-wise Accuracy36.6
4
Showing 8 of 8 rows