I don't really disagree, but that would make for a rather perverse unit, given that inverse temperatures are already on a logarithm scale relative to the probabilities of microscopic states.