Add simple interpretation assistance to metric docstrings

[leanderweber]

Description of the problem & solution

Currently, the docstring of each metric mostly explains what the metric computes, as well as the corresponding reference. It may be helpful in practice to add a very simple guide for how to interpret the resulting scores as well. E.g., just something like "Higher/Lower scores are better". I believe this would increase usability a lot, especially for users that are new to XAI quantification.

Minimum acceptance criteria

• Add guidance for how outputs should be interpreted to each metric docstring,
• @annahedstroem