Maybe I am a bit confused but I just meant to imply that A and B could be considered equally probable if you're a Bayesian who has absolute confidence that she's the kind of person who is active in the feminist movement.
100% now is 100% forever, in the Bayesian framework.
If a Bayesian agent is absolutely certain that she is active in the feminist movement, it would mean that absolutely no evidence could convince it of the contrary. Even if we learn that she was kidnapped and forced to work in, say, Saudi Arabia as a bank teller where she's forbidden from feminism, even then, the Bayesian agent would have to still stay at 100% confidence that she's active in feminism.
How could A possibly be less likely than B?? The occurrences of B are a strict subset of A, there is no possible way B could happen more frequently.