The field descriptions for Float.NaN and Double.NaN use the term "equivalence", but for a reader not well-versed in floating point, it may not be immediately clear what "equivalent" means in this context.

A simple way of improving this would be to add a link to the equivalence discussion in Double