Which two features achieved the highest cross-validation accuracy when used alone in the model?

Which two features achieved the highest cross-validation accuracy when used alone in the model?: a task in SmolDataEnvs-harbor-train (Harbor dataset). Files (in /home/user/input, no subfolders): - Iris.csv