Postdoc application: Should I mention who I'd like to work with?

I'd suggest that you don't name anyone unless you somehow name them all. Not only do you avoid seeming inflexible, you actually maintain some flexibility until you can judge better who would be best if you need to single out anyone.

Indicate that you are familiar with their work somehow, I think, by naming papers you are familiar with. (Familiar with, not just saw somewhere).

Truly collaborative groups probably think of themselves as a unit in some ways.

On the other hand, mentioning specific topics that most interest you would seem like a better option. That might imply one or the other of the members, of course, without stating it.