Returns largest integer <= the base 2 log of the magnitude of ARG. This is the same as the exponent of a float. (fn ARG)