This trend has continued more recently with the power increase.
Figure 1 shows the spectrum:
* blue - back in September when the range was ~30Mpc
* red - after the SDB1/SPRB intervention and the 1/f^4 noise (mirror actuator saturation) fixed
* green - with input power increased from 14W to 20W, not much change compared to red
* black - with input power of 26W, flat noise much higher
* magenta - same data as blue (Sep 27) + a flat noise at 1.9e-23m, matches very well the black curve
So the mystery noise is still very consistent with flat noise and might be increasing with input power.
