In case it's not obvious, you're generally going to end up with unrealistic stars if haphazardly changing the appmag value, for precisely this reason. If you're telling SE to make an M type star with an apparent magnitude of -50 (which is a billion times brighter than the Sun by the way -- Earth would be vaporized!), but many light years from Earth, then of course it is going to be absurdly huge.
You would be better off working in absolute magnitude, using an established range of abs mags that stars of a given spectral class can have according to an HR diagram.
Oh, fully aware and definitely know the settings I have are pretty unrealistic (maybe not in the grand scheme of things, but definitely from what we know of our universe)! I definitely agree completely with the purpose of apparent magnitude.
I decided to toy with what you all have said about negative values and found (based on the M9 0) that after around -14.56 +/- that the star becomes an orange elliptical with no physical model. If I go to Earth and view that same star it will cloud the entirety of one part of the milky way in insane amounts of light.
This is why I love this sandbox so much. Little tweaks and seeing how they generate in here is so much fun for me.
Recently I put a black hole within 600K km from the SUN and it was entertaining to see it from other planets.