WangResearchLab/verification-evaluation-data
收藏Hugging Face2025-10-17 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/WangResearchLab/verification-evaluation-data
下载链接
链接失效反馈官方服务：
资源简介：
这是一个包含数学问题及其解答的数据集，分为aime2024、aime2025、livebench_math、gpqa和validation五个部分。每个问题都有问题描述(prompt)、标准答案(gold_standard_solution)以及模型生成的多个响应(llm_responses)，还包括一个等价矩阵(equiv_matrix)。数据集总大小为1.87GB，下载大小为683MB。
This is a dataset containing math problems and their solutions, divided into five parts: aime2024, aime2025, livebench_math, gpqa, and validation. Each problem includes a problem description (prompt), a standard solution (gold_standard_solution), multiple responses generated by the model (llm_responses), and an equivalence matrix (equiv_matrix). The total size of the dataset is 1.87GB, and the download size is 683MB.
提供机构：
WangResearchLab
