What are the 'support vectors' for SVM?

a) Data points that are crucial for defining the decision boundary.
b) Vectors representing the features in the dataset.
c) Points with the highest dimensionality.
d) Points with the lowest dimensionality.