I actually was going to link the same one I always do, which I think I heard about through a blog or talk. If that's not good enough, it's easy to devise your own test and put it to an LLM. The way you phrased that makes it sound like you're more interested in ignoring any empirical evidence, though.