I would be inclined to keep the integrity of the histogram intact and
plot a second one (perhaps side by side) with just those bins that are
obscured by the scale of the first graph.
It does beg the question of whether those bins are important in the
context of a bin or bins with such large frequencies.
Greg