MM-A
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MM-A out-of-distribution (OOD)
90Task Accuracy
6
MM-A in-distribution (test)
98.9Accuracy
6
MM-A Extrapolation 5 digits
44.06Task Accuracy
5
MM-A Extrapolation 4 digits
69.73Task Accuracy
5
MM-A (out-of-distribution)
65.46Concept Accuracy
2
MM-A in-distribution (test)
85.16Concept Accuracy
2