Color scheme in tables doesn't respect metrics ranges.

[g8a9]

When we visualize faithfulness and plausibility metrics we use a color scheme where, generally, the darker the cell, the better.
However, color ranges are unclear and seem to be different for faithfulness and plausibility metrics.

E.g., see below

Here aopc_suff has the worst value with 1 and is correctly displayed in white, but plausibility metrics based on F1 scores (e.g., token_f1_plau) are not shown in white when they are 0 (worst value).

I think we should match the color and metric ranges, and also add an extensive user guide in the doc to clarify the matter.