It seems there is something odd happening, It's always a concern if you can't identify what is really wrong and keep taking pot shots at it as you don't know if you've really fixed the problem.
Could be focusing on the wrong angle for a reason why, may be worth describing the logic to see if anyone can spot a prob in it.
If the cam cradle plugs are not leaking (they go into clear air and not the gasket anyway, don't they) the only scource of pressurized oil is from the O-rings protecting oil supply to the cam area, so the oil must be coming from them as a primary fault, right or wrong?
The head gasket is primarily there to seal the barrel to combustion chamber area and the rest is to take up any minor variation in the machined faces, if it was the original intention to stop pressured oil it would probably have had something like the barrel rings in copper around oil ways. Back to the O-rings?
Why don't they seal? If the O-rings were too big (total volume) they would hit the capacity of the counterbore and possibly prevent the head properly compressing the barrel rings, so leaking compression?
When the O-rings have been in a long time they look square is section showing that they've been squashed and seal against the casting.
Soooo the problem appears to be that the rings are not sealing against top / bottom faces, Why? (sounds like a corporate problem statement!! bit of blue sky, out of the box thinking there
)
What affects that:-
Dimension, is it correct?
Material, some rings are noticeably harder that other (pattern? some applications would have a different material spec to this one, think it's called Sure rating) and would affect the physical sealing to the faces.
Clamping, assuming this is correct, but are you sure of your torque wrench accuracy. I've always used a small scale one for this 0-40lbs as it gives good seperation of scale and more accurate actuation at light loads. Have you got any genuine new Honda O-rings to use?
Hope this is of some use, any comments welcome.