Its most likely a combination of float and overlap...overlap will do the same thing to vacuum readings once rpm's get high enough, if the overlap is biased towards the intake event...above 6,000 or so rpm, you're literally talking about 1 or 2 degrees too much overlap and the above is the result...
So, before you order valve springs...i would try to move that overlap (keep the same amount) a little 'earlier' in the event (remember, 'later' would be biasing towards the intake event, as the exhaust stroke is the beginning in terms of overlap)...by a degree or two to start...you may even that out surprisingly well...
my old engine build hit those numbers because of revs...I really don't even think the compression played a huge role in the peak hp numbers, because dynamic compression was down due to overlap...but i had tuned in literally as much overlap as i could, which was very exhaust stroke sided...so my intake vacuum remained strong, and exhaust scavenging was great at higher rpm...
also, for what its worth...that engine set up would hardly be what you're after...idle was non-existent, and mid range torque simply wasn't there...unless you could keep the car on boil above 5000 rpm in all track conditions, it would've never made a protege 'fast'...
I ran Eiback valve springs in that head, with stock everything else (and pretty much twiggy cams, slightly different as it was my own design...but its what Andy and i worked on together before the group buy)...but i wasn't seeing any crazy float before i put the springs in...and that included pulls up to around 7600 rpm...so i'm guessing you may have a combination of the two...and changing where the overlap peaks can drastically effect intake charge...i bet you could dial out a good deal of your problem, and not even need to worry about springs right away...with the overlap dialed in, you'll be quickly sitting around 170whp i'd bet...
If there is float, though, it definitely isn't a lot of it...valve float will cut power nearly as bad as straight up pre-ignition/denation/missing, and the engine will struggle to keep climbing in revs because of it...you're holding your power pretty well for nearly 1000 rpm up there, and still accelerating the bottom end...i think its the overlap bleeding a fixed amount of pressure off at each cycle that is more of the issue here...