michaelcohen comments on Formal Solution to the Inner Alignment Problem

[michaelcohen]

Does it make sense to talk about ${\check{\mu }}_1$, which is like ${\mu }_1$ in being treacherous, but is uses the true model ${\mu }_0$ instead of the honest model ${\hat{\mu }}_0$? I guess you would expect ${\check{\mu }}_1$ to have a lower posterior than ${\mu }_0$?