Fixed raising integer to negative numpy integers

[jbloom]

Previously this error was sometimes raised in breaks if a base was being raised to a numpy array that included negative integers: https://stackoverflow.com/questions/43287311/why-cant-i-raise-to-a-negative-power-in-numpy

Fixed this by making the dtype of the array that is the exponent a float.

[has2k1]

Thanks. I had caught this and started fixing it just after making the previous release.