LLM Olympiad: Why Model Evaluation Needs a Sealed Exam
This addresses the issue of misleading progress metrics in NLP for researchers and practitioners, though it is incremental as it builds on existing closed benchmark concepts.
The paper tackles the problem of unreliable LLM benchmarks by proposing an Olympiad-style evaluation with sealed problems and frozen submissions to prevent gaming, aiming to make strong performance harder to manufacture and easier to trust.
Benchmarks and leaderboards are how NLP most often communicates progress, but in the LLM era they are increasingly easy to misread. Scores can reflect benchmark-chasing, hidden evaluation choices, or accidental exposure to test content -- not just broad capability. Closed benchmarks delay some of these issues, but reduce transparency and make it harder for the community to learn from results. We argue for a complementary practice: an Olympiad-style evaluation event where problems are sealed until evaluation, submissions are frozen in advance, and all entries run through one standardized harness. After scoring, the full task set and evaluation code are released so results can be reproduced and audited. This design aims to make strong performance harder to ``manufacture'' and easier to trust.