Logarithmic spectrum.

I dislike the concept of a single-axis, back-and-forth "spectrum" such as your 12% would imply, but more of a multidimensional conglomerate of myriad spectra (plural), where e.g. spoken language when not overstimulated is one, spoken language ability WHEN overstimulated is another, social interactions when not oversocial-ed is one axis, social interaction ability while oversocial-ed is another, tactile sensory is one axis, smell/olfactory sensory is another, visual sensory is another, etc.

Because that's the reality, is that it's a mish-mash of several interconnected axes and spectra, that tend to go off the rails a bit faster / more easily than NT / allists seem to go off the rails and break down