Nice impressions! They sound like something I would like actually, but just have no real need right now.
There's some confusion or misinformation or translation error, possibly from the manufacturer side for the D100, but from my and another head-fier's experience, there might be some sort of resistor in-line to reduce the output of the low jack (likely not present on the high jack) which likely actually increases the output impedance significantly changing the FR on IEMs with non-flat impedance curves and low input impedance.
Per Rins measurements:
You'll note using Rins standard compensation that there is a reasonably significant drop off after 10k as you noted and the peak at 7k is only really a peak due to the adjacent dips, but nonetheless present as is. And that with 33+ ohm added impedance, you'll see a ~10dB drop off of pretty much everything above 2KHz.
But it's worth mentioning that a new compensation method that is gaining traction, Olive-Welti shows a slightly different story for the treble:
Though I think that for IEMs this compensation shows just a weeeee bit more of a subbass drop than I usually hear. But that's just me.
Most reviewers of the D100 probably only did extensive testing with cans rather than IEMs which is why this catches IEM users by surprise.