I feel that two things may be happening. Especially since this effect was noticed on the early mounts done in the patination process, I'd venture to say that the initial surface prep before immersion in the patina solution was probably a bit uneven. Also, trying to get a dark coloration by keeping the brass in the solution for long periods can definitely lead to the purpling you are seeing, and can go all the way to silvering. I would recommend re-scuffing with fine scotch-brite and doing a series of shorter dips with washes in between to try to achieve the color you want.

The baking soda solution does make some sense as the patina solution contains an acid like Selenic acid and the baking soda would be acting to neutralize it. It would be interesting to do some experiments with and without a soda bath to see if there are different outcomes over time. I look forward to seeing what you discover.