I can not answer about the deadliness of GRBs, but I think you will find those answers in Phil Plaits book "Death from the Skies!".
- How many civilizations might form on bodies with very thick atmospheres, far from their Suns? (Venus does not need a ozone layer to keep the UV out, and might be very habitable a few AU out.)
Yes, insulation is a good idea. But the planet will always radiate as a black body and loose energy, which has to be re-supplied by the suns radiation. The radiation drops with the square of the distance, so rather quickly. These considerations (make-up and size of planets) go into calculations for the habitable zone.
I can also imagine that a GRB comes with considerable photon pressure and might strip the entire atmosphere off a planet, or heat it to a point where it dissipates into space.
- How many planets might have very long rotation periods (years), so that the night hemisphere never is subjected to the daytime UV?
I think the rotation of planets around their own axis (spin) is not known outside the solar system. Generally, the spin is generated from formation of planets in the rotating protostellar disk, but interactions and changing orbits may modify the spin (Venus, Uranus).
- Are there rotation axis directions and orbital precession constants for planets that would keep GRB radiation mostly in one hemisphere, leaving the other to develop?
If you do not have the problem of heating and evaporation of the atmosphere I mentioned above, then yes, that is probably possible. For example if the GRB goes off from the direction of the spin axis ("below/above the solar system"). This may safe you from one GRB, but since GRBs come randomly from all directions it is not failsafe across many billion years.
- How many planets might have other special circumstances that protect their ozone (such as a lack of N2 in their atmosphere, or an ozone generating biology in their stratosphere, etc.)
Not sure. I think it is possible to come up with such scenarios as you stated, but it has to be shown that they are frequent occurrences to be relevant for changing the survival rate of complex life.