Rubber mallets have failed more often than succeeded with jobs like this. Maybe they've never succeeded?
You need a "shock load" as well as any direct force that you're applying with the hammer. The shock deforms the metal microscopically and the vibrations are a key to helping get the job done. Again, not just force.
But leave the puller in place and under tension. This usually speeds up any hammering process.
I've always hesitated to use either a pickle fork or a single hammer on a steering box. On a tie-rod end, not a problem. But on a steering sector shaft in bushings/bearings, was never sure. So I use the 2-hammer method.
That's where one hammer (if different, the heavier one) is laying up along side the arm, and the second hammer is doing the hitting. This does two things. One is it gives sort of a resonating second shock wave and two, takes some of the sideways force out of the blow from the other hammer.
I try to angle the puller in such a way as I can get the hammers to lay up against the main large portion of the arm. That's not always possible of course, and your puller type does take up some real estate just sitting there. But do what you can.
In a pinch, you can remove the puller, smack the crapola out of the arm a few times, then re-apply pressure from the puller. The shock waves might have done their job and the arm will come off easier after. Done that a time or two myself.
Basically you do what you have to do, but use a hammer at least on some attempts because it really works on tapered stuff.
Good luck.
Paul