net.IW{1}
will give you the codebook vectors/weights. For example:
x = simplecluster_dataset;
net = selforgmap([8 8]);
net = train( net, x );
net.IW{1}
It isn't the most intuitive to find the things you want in an SOM from the Deep Learning toolbox. I constantly forgot where all my parameters were when I used it. It uses the same network class as deep networks so there is a lot of stuff in there totally irrelevant to SOMs.