Two equally successful statements of purpose can look very different from one another. Some graduate students enter with broad interests and others enter already knowing a specific question they want to work on. Both types of students can be good applicants and will have very different personal statements, potentially with very different levels detail when describing past research. Your advisor will give you much better, and more specific advise than the general advice you would get from academia.stackexchange.
Write a draft of your Personal Statement and then show it to your advisor [or other math professor who knows you well] to get feedback. After editing it based on this feedback, it is also a good idea to ask for feedback from professor(s) you intend on asking for a letter of recommendation. This serves two purposes (1) likely these letter writers will be less knowledgeable of your research area and give you a different perspective than your advisor (one that may mimic the academic diversity, to some degree, on an admissions committee) (2) Them giving you feedback on your Personal Statement will allow them to get to know you better and hence write you a better reference letter.
Generally speaking, if you can go into some detail, this is a good thing, and will set you apart from other applicants. In pure math, it is quite rare for students to have a very specific research direction before graduate school. Of course don't go so far into detail that only specialists can understand it.