The hot water freezing experiment is not the same thing, as this is due to evaporation resulting in less water there to freeze by the time it does.
I would expect the gas ram to heat by a similar amount, but maybe the cooling of the action gives a much larger heat sink/temperature differential for the ram to cool again after heating? The pressure in the ram will certainly increase with temperature, but friction of the seals and relative expansion/contraction of metal parts will all play a role.