OK, some more testing and I think I got it, thanks to a certain Gaussian function ;P
Now I'm trying to figure out how to clamp and get actual values of light distance from -1 to +2 so that the dilation is light dependent, but am unable to get it right, the eye just...wobbles in place when I put up the light. As it shows in the image, the light wobbles the iris on the RG axis, and doesn't actually spread linearly.
If anyone has tips about this, I would like to hear it.
EDIT: OK, a cosine/sine function with a linear clamp works wonder for the dilation amount, but I'm still getting the wobble and weird overlay issue previously mentioned.