LLM Olympiad: Why Model Evaluation Needs a Sealed Exam
arXiv:2603.23292v1 Announce Type: new Abstract: Benchmarks and leaderboards are how NLP most often communicates progress, but in the LLM era they are increasingly easy to …
Jan Christian Blaise Cruz, Alham Fikri Aji
2 views
