When Benchmarks are Targets: Revealing the Sensitivity of Large Language Model Leaderboards

Alzahrani, Norah, Alyahya, Hisham Abdullah, Alnumay, Yazeed · arXiv · 2024