But where do those 0.04/TWh deaths in nuclear come from? Are they mostly from a low-probability high-death count event? Is it 1 person per year over a thousand years from solar, or maybe the whole village at once in a thousand years, from nuclear?

If it really is more along the lines of nuclear is safe, but once every five years a worker at the plant might slip up and be exposed to dangerous levels by accident, then I guess I think the comparison makes more sense.