Decision models like Jev don't beat LLM-as-a-judge or traditional classifiers

prompts that these evaluations were done are too trivial

[flagged]