What are you looking for that isn't simply a math textbook? Honestly, that's how mathematicians learn "actual" mathematical subjects, too.
If you're comfortable with calculus as a subject, for example, and want a "pure mathematics" approach, I recommend Michael Spivak's Calculus. If you've never worked through a pure math textbook from start to finish, that's a good start.
There's not that much interesting in the three subjects you listed — logic, set theory, and category theory — that doesn't depend on other subjects or a prior level of mathematical maturity. Category theory was originally invented to solve and categorize problems in algebraic topology, for example.
For example, I rather like mathematical logic and model theory, but you're going to have a rough time if you don't have a visceral, intuitive understanding of countability arguments and at least a handful of subjects you'd be reasoning "about." Unless you know the standard model of arithmetic, for example, how can you think about non-standard models? Without that, important results like the Löwenheim–Skolem theorem will likely seem contextless.