I think the temperature zero or a fixed seed are not a blocker for this expiment if you sample multiple values and compare the distributions.
I think the temperature zero or a fixed seed are not a blocker for this expiment if you sample multiple values and compare the distributions.