There are several topics about the display driver crashes that don't appear to be hardware related and I'd just like to post what fixed the problem for me.
Problem:
The problem appears to have something to do with the downclocking/upclocking procedure OR the 2D clock speeds. I believe this because the solutions below fixed the problem and if you do not install drivers (which enable the downclocking) and just use the Standard VGA controller windows installs, the problem is non-existant.
First Solution: Force 3d performance levels via Rivatuner (this solution is noted in a different post--details
here)
Using drivers 186.18 and Rivatuner, by forcing 3d peformance levels I haven't had a single crash since. The con to this method is much higher idle heat and power consumption.
Second Solution: Increase 2d clockspeeds
I don't prefer the first solution because of the increased idle temps. The GTX 260 has a core clock of 300 and memory of 100 at 2d clock speeds.
I originally clocked them to 400/150 and experienced no crashes over 24h. I then experimented and began downclocking to see how far I could go. Initially, I decreased both speeds simultaneously; crashes began to happen around 350/115.
Then, I decided to just clock the memory. At 300/150 I had greatly reduced crashes (compared to 300/100), but still some over 24h. Right now I'm at 335/150 and have had no crashes in the last 24h.
Conclusion -- It seems the memory clock speeds has more bearing on the driver crashes than the core clock.
Unfortunately, because of the intermittent, random nature of the crashes, it is very difficult to robustly test these methods. My "stability testing" was simply any computer use over 24 hours since there are no triggers to the crash. Remember, systems experiencing this problem seem MOST stable during 3d clock speeds and rarely crash during game-play, but more frequently crash during 2d clockspeed times.
Hope this helps someone.