Documentation of Unit.format() is incomplete

**Describe the bug**
The documentation at https://mathjs.org/docs/datatypes/units.html just says "The function will determine the best fitting prefix for the unit", and the Class page at https://mathjs.org/docs/reference/classes/unit.html. One place or the other should provide detail on how the best prefix is selecting, documenting the bias toward the current unit, and the preference for a larger number of tenths over a large number of hundreds (which not everyone may agree with). The absence of such documentation leads to concerns such as #3430.