Talk-Polywell.org

Looking at a cross section of the Kiwi engine, http://up-ship.com/blog/?p=24791, got me thinking of a variant. Instead of running the hydrogen propellant through the graphite moderator then the nuclear core, run it backwards. My novice level materials tech tells me the graphite can handle higher temperatures than the fission core. The result could be the core heating to its limit, then the reflector, heated by nuclear radiation, giving a modest additional heat.

As I recall off the top of my head, the point is for the propellant to cool the moderator, and get some pre-heat at the same time.
Doing it backwards would defeat that.