A "sock puppet" account is one which is literally controlled by the same person, like a puppet on your hand. So for example, if I made a second account called "John Gammon" (using a slightly different email address) and then whenever I made a post I got my "John Gammon" account to upvote it, that would be a sock puppet vote. Clearly that would be unethical, as no person, other than myself, was voting for it.
In that case, I would hardly need to get permission from myself to do the upvote, as it was me doing it.
However if a different person, Joe Bloggs, just happens to think I am wonderful and vote all my posts up, then this isn't a sock puppet vote. That is just an admirer. The Stack Exchange algorithm for detecting sock puppet votes might not be able to distinguish between the two, however.
Please do not vote for my answers. I help people because I want to, not for the reputation
That is certainly admirable to want to help people, however the whole idea behind Stack Exchange is that good answers are voted up, and bad ones voted down. Asking people to not vote up your answers, if they are good, defeats that mechanism.
Next, you'll be asking people to not accept your answers, because that also gives you reputation. However accepting answers indicates that your answer solves their problem.
The up-vote is not so much to give you reputation (that is a side-effect) but to mark the answer as good and useful.
Who was the “deleted” user who was responsible?
I don't know the answer to that. Not even moderators can find that out. I suggest, though, that new users don't just go around voting hundreds of posts up willy-nilly. It might look to "the algorithm" that they are sock puppet votes.
You tagged this question "grow-up-se-rep-is-worthless". It isn't worthless — it is the basis of the entire system. Without reputation, you may as well have the to-and-fro of a forum. With a forum, it is hard to see which are the good answers and which are the rubbish ones. The reputation system (and the voting) is specifically designed to let third parties know whether they should trust one answer over another one.