Some time back we had two candidates for a backend role, and their take-home tasks came back almost identical in quality. Clean, tested, both fine. So we gave each of them the same follow-up: what would you remove from your own solution?
One said nothing. Every piece had a reason, and they could defend all of them. The other crossed out two of their own files on the spot, and explained that those were there out of habit, not because anything needed them.
We hired the second one. The first answer was not wrong, but we had already paid for that habit. We have inherited systems where every feature was built to handle a future that never arrived, and the first months of that kind of engagement go into reading code nobody needs anymore, before you are allowed to change anything useful.
So "what would you cut?" is now a standard question for us, for engineers and for leads. Building is teachable. Knowing what does not need to exist usually is not: people either learned it from maintaining their own work for a few years, or they have not yet.
You probably have your own version of that question, the one that tells you more than the test itself.