Notes on logical priors from the MIRI workshop — AI Alignment Forum