Hi Ryant, which specific page are you referring to and what MATLAB release are you using?
For the iris_dataset data, yes you are right, irisTargets is a 3x150 matrix of associated class vectors defining which of the three classes each input is assigned to, as documented here:
where you can find:
Examples
Classify Using Softmax Layer
Load the sample data.
[X,T] = iris_dataset;
X is a 4x150 matrix of four attributes of iris flowers: Sepal length, sepal width, petal length, petal width.
T is a 3x150 matrix of associated class vectors defining which of the three classes each input is assigned to. Each row corresponds to a dummy variable representing one of the iris species (classes). In each column, a 1 in one of the three rows represents the class that particular sample (observation or example) belongs to. There is a zero in the rows for the other classes that the observation does not belong to.
