Passing the Test the Language Models Failed

The strictest version of the realization check asks three independent kinds of evidence to agree. The language models scored zero. Here's what happened when I ran a continuous carrier through it — and why passing turned out to be the opposite of automatic.