# SVM's support vectors decision function representataion

I am currently using SVM for my project with 'rbf' kernel.

What i understand from the theory is that the decision function value for the support vectors must be either +1 or -1. (used clf.decision_function(x))

But i find some support vectors decision function value is even 0.76, -0.96, 0.93 and so on.. (its not even closer to +1 or -1).

What is wrong in this scenario, or whether my understanding is wrong btw ?