If you have a pure DFS for your state-expansion strategy, then your agenda works like a stack, holding at most n partial solutions, the current state you are examining, and all ancestor states. Should a computational path find a new bssf, then all states on the agenda are ancestor problems of that bssf and cannot be pruned.
Given the top partial solution on your stack, you may decide not to explore some of it's possible child states because of your bounding function and bssf. That decision to not explore those computational sub-trees is what we mean by implicit pruning. The number of such states is not reported. If they were, they would need to be displayed in scientific notation for the largeness of them. It is okay to report a 0 for number of pruned states if that is part of your strategy.
Chris