Kevin - can you confirm (or refute) the idea that reheating O-1 to non-magnetic after quenching "wipes the slate clean"?
(respectfully quoting a post above)
Daniel, as with most of this stuff, there are no absolutes, so it depends upon what areas you look at. In the terms of prior austenite grains and the new martensitic structure, it will begin recreating itself as soon as you approach critical. Carbides, even simple cementite (iron carbide) requires more effort to wipe clean , that is why there is an Accm line on the iron carbon equilibrium diagram.:
on the diagram there are three lines to the far left, A1, A2, and A3, since we are talking about heating so they could also be called Ac1, Ac2, Ac3. A1 is the point at which the transformation (alpha iron to gamma iron) begins. Before this you get degrees of tempering carbide formation and then formation of spheroidal cementite in increasing degrees. As you reach Ac1 the pathways for the carbon to move begin to open up and you get much more diffusion and as more gamma iron is formed the carbon is dissolved in it. Hanging around or just below Ac1 can allow the carbon to move, but to find its own kind and pool up and segregate, this is how spheroidizing works. If there are carbide forming elements that can grab onto that carbon they will gorge on it and this is how you get banding and heavy segregation.
Continue to heat the steel and you will reach Ac2, the Currie Point or nonmagnetic, which, believe it or not, is not all that significant for this conversation. But now since we are talking about carbides lets jump to the other side of the V on the chart, the portion that rest between .8% and 2% carbon and would describe a higher carbon steel. The next line up is Acm. Between Ac1 and Acm is a whole field of varying degrees of solution for simple cementite, stop at any point in there and you will have undissolved carbide that is leftover from the last heat, pass the Acm line in temperature and you will dissolve it all and have all fresh austenite, unless
Unless there are carbide formers, then you will need to soak for an extended period to peel some extra carbon off them or increase the temperature high enough to break the chemical bonds ( for something like vanadium this can be in excess of 1900F.)
Are you starting to see how this is not so black and white?
The first thing to be wiped clean on a reheat is residual strain from forging, cold working, thermal treatments etc. In a process known as recovery the dislocations will be rearranged in a configuration more conducive to lower stored energy, this is how stress relieving works and why the idea of stresses or strengthening from previous deformation, be it hot or cold, surviving through a heat treatment is nonsense.
The next thing to go (I am probably missing some) would be the original even grain configuration as new grains begin to nucleate and from sub grains at the old grain boundaries, you would now have a mixed grain size between the remnants of the old grains and the new. When the process is complete there will be fresh austenite grains that will then set to consuming any extras carbon it the grain boundaries as heat increases. Grain growth cannot occur until those boundaries are depleted and destabilized, above Acm. But if there are alloy carbides they will arrest this and that is why that 5 hour soak was possible.
Finally the last thing to go, and would probably survive most of our heat treating operations would be those richer alloying carbides.
So, in short,

the answer to your question is YES! ... and NO!