Difference in unstructured pruning with L1 vs L2 norm

Question as simple as the title. What is the difference in results between L1Unstructured and LnUnstructured when using the torch.nn.utils.prune class. If say, I pruned 20% a dense layer’s weights with L1 and L2 norm, respectively, wouldn’t both simply prune the lowest 20% of the weight elements in absolute value?