Pruning is setting weight values to 0. Simple as that. If you set enough of them to 0 then it might be worthwhile converting the weights to sparse matrices, my software can do that. This is probably best shown in a notebook. Mostly I don’t convert to sparse neural networks, there are a lot of other reasons to prune models.
I have shown that you can modify behaviour without training. You can make a model smaller, and more efficient. You can even seem to get small increases in test accuracy. As for the how, you can read the code, you can use my code, you can work out how to do it in your chosen software, it is not that hard.