Large Language ModelsOpenProblemBench is Prepared to Test Scientific Limits of Artificial Intelligence
The new benchmark published in ArXiv measures the ability to make real scientific research of machines by offering 82 undissolved problems of math and
