Saying that tracking model growth directly populates a dataset feels like a big stretch, frankly.
In my line of work, we track RSVPs for a wedding, but that list doesn't magically make the guests exist or decide to show up.
The models grow, that's the independent thing, and then the tracking just records that existing expansion.
It's a measurement, a logbook, not the engine making those thousands of models appear.
If you confuse the two, you miss who's actually building things, like a caterer thinking their menu creates the food, instead of the ingredients.