I recently edited an old and very negatively voted question (-11) in order to improve it and turn it into a not so bad or maybe even good question. However, a big problem that I see is, that while users who once downvoted it might reconsider their voting decision, they mostly just won't notice the change at all, since such a negatively voted question is actually hidden from the normal site frontpage (which only shows questions >-4, >-8 on meta), which is still what (I'd admittedly speculate) a large majority of users uses as their primary question feed.
Now there are certain means to make other users more aware of the situation, like putting it into chat. But another option I saw as a nice idea (especially when the question doesn't have many good answers already, even though that wasn't the case in my example) was to put a bounty on it. But alas, even with the bounty it was still hidden from the frontpage.
Now I see that the hiding of low quality questions from the site frontpage is not a bad idea. However, I think significantly revising an old bad question is a valid case, even if maybe a rare one, where we want the additional attention on the question. Of course one could argue that people could circumvent the hiding of "genuine" low-quality questions this way, which should even be hidden albeit having a bounty. But a bounty comes with a certain effort and enagegment to the question that I think makes it important enough to override the automatic hiding of low quality questions. Afterall we want bountied questions to gain as much attention as possible, which is the primary use case of bounties.
So I would propose that bountied questions are not hidden from the frontpage disregarding their possibly very negative votes.
I know that the whole navigation and frontpage are currently under significant revision. However, the new navigation still seems to hide very low-score questions to some degree. So I think this feature reuqest is largely orthogonal thereto. But maybe this whole "question hiding feature" gets scratched in the future, which would make this feature request obsolete.