"Model A wins on MMLU. Model B wins on ARC-Challenge. Model C wins on HellaSwag.
At some point you stop trusting any of them—not because benchmarks are meaningless, but because no two of them seem to tell the same story about which model is actually better.
[…]
We found a fix. It’s called Train-before-Test."
