OlympiadBench
收藏资源简介:
OlympiadBench是一个高难度的双语多模态科学基准数据集,由清华大学计算机科学与技术系人工智能研究所创建。该数据集包含8952个来自国际奥林匹克竞赛、中国奥林匹克竞赛和中国高考的数学和物理问题。每个问题都配有专家级的逐步推理注释,旨在评估和推动大型模型在数学和物理推理方面的能力。数据集涵盖了广泛的难度和问题类型,适用于评估和提升模型在复杂科学问题解决方面的性能。
OlympiadBench is a high-difficulty bilingual multimodal scientific benchmark dataset created by the Institute of Artificial Intelligence, Department of Computer Science and Technology, Tsinghua University. This dataset contains 8,952 math and physics problems sourced from International Olympiads, Chinese Olympiads, and China's National College Entrance Examination (Gaokao). Each problem is accompanied by expert-level step-by-step reasoning annotations, aiming to evaluate and advance the capabilities of large models in mathematical and physical reasoning. The dataset covers a broad spectrum of difficulty levels and problem types, and is suitable for evaluating and enhancing the performance of models in solving complex scientific problems.

- 1OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems清华大学计算机科学与技术系人工智能研究所 · 2024年



