Weird simulation results in latest Spectre Transient Noise Simulation RAK

While reading the RAK titled "Spectre Transient Noise Simulation from ADE - Introduction" and trying out the related testbenches, I stumbled upon a simulation result that looks quite strange to me.

In the first tutorial, the schematic is very simple: a noisy resistor driven by two current sources: a 1 μA DC source and a sine source with frequency 500 MHz and amplitude 1 μA.

On page 19 of the RAK, the plot of the power spectral density of the signal (without noise) shows many harmonics. This is more evident in Figure 15 (page 22). The text even states that "The signal spectrum has three peaks at 500 MHz, 1.5 GHz and 2.5 GHz as expected". Now, this is not what I would expect. I would expect the spectrum to show a DC component and a single peak at 500 MHz, since we are dealing with a simple sinewave with a DC offset. Am I missing something here?

The strange thing is that in the next tutorial of the RAK (which uses the same circuit) on page 31, you can clearly see that the harmonics have disappeared from the spectrum. I looked for differences in the netlists, but couldn't find any.

What's even stranger is that in an older version of a similar document (titled “Virtuoso Spectre Transient Noise Analysis - product version Spectre 18.1”), the same first tutorial shows no harmonics, even though it uses the same testbench.

Does anyone have any idea what might be causing these harmonics in the first tutorial of the new RAK?