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

TableShift

Benchmarks

Task NameDataset NameSOTA ResultTrend
ClassificationTableShift Sepsis OOD v1 (test)
AUPRC @ R<0.182.85
14
ClassificationTableShift Hospital Readmission OOD v1 (test)
AUPRC@R<0.190.52
14
ClassificationTableShift FICO HELOC v1 (OOD test)
AUPRC (R<0.1)96.99
14
ClassificationTableShift Childhood Lead v1 (OOD test)
AUPRC @ R<0.199.72
14
Tabular ClassificationTableShift Acspubcov (OOD)
Accuracy70.6
3
Tabular ClassificationTableShift Acspubcov (IID)
Accuracy80.4
3
Tabular ClassificationTableShift Acsincome IID
Accuracy77.5
3
Tabular ClassificationTableShift-Diabetes IID
Accuracy64.9
3
Showing 8 of 8 rows