Beyond the Leaderboard: Rethinking Medical Benchmarks for Large Language Models | Read Paper on Bytez