taowangcheng/Original-MMLU
收藏资源简介:
--- dataset_info: - config_name: benchmark features: - name: question dtype: string - name: subject dtype: string - name: choices sequence: string - name: answer dtype: class_label: names: '0': A '1': B '2': C '3': D splits: - name: test num_bytes: 6967453 num_examples: 14042 download_size: 3452750 dataset_size: 6967453 - config_name: train features: - name: question dtype: string - name: subject dtype: string - name: choices sequence: string - name: answer dtype: class_label: names: '0': A '1': B '2': C '3': D splits: - name: full num_bytes: 763484 num_examples: 1531 - name: STEM num_bytes: 160077.3115610712 num_examples: 321 - name: humanities num_bytes: 258317.90463749183 num_examples: 518 - name: social_sciences num_bytes: 168056.24297844546 num_examples: 337 - name: other num_bytes: 177032.54082299152 num_examples: 355 download_size: 813056 dataset_size: 1526968.0 configs: - config_name: benchmark data_files: - split: test path: benchmark/test-* - config_name: train data_files: - split: full path: train/full-* - split: STEM path: train/STEM-* - split: humanities path: train/humanities-* - split: social_sciences path: train/social_sciences-* - split: other path: train/other-* ---
数据集信息: - 配置名称:基准(benchmark) 特征项: - 名称:问题,数据类型:字符串(string) - 名称:学科,数据类型:字符串(string) - 名称:选项,数据类型:字符串序列(sequence of string) - 名称:答案,数据类型:类别标签(class_label),其类别映射为:'0'对应A,'1'对应B,'2'对应C,'3'对应D 数据集划分: - 划分名称:测试集(test),字节大小:6967453,样本数量:14042 下载大小:3452750 数据集总大小:6967453 - 配置名称:训练(train) 特征项: - 名称:问题,数据类型:字符串(string) - 名称:学科,数据类型:字符串(string) - 名称:选项,数据类型:字符串序列(sequence of string) - 名称:答案,数据类型:类别标签(class_label),其类别映射为:'0'对应A,'1'对应B,'2'对应C,'3'对应D 数据集划分: - 划分名称:全量集(full),字节大小:763484,样本数量:1531 - 划分名称:理工科(STEM),字节大小:160077.3115610712,样本数量:321 - 划分名称:人文学科(humanities),字节大小:258317.90463749183,样本数量:518 - 划分名称:社会科学(social_sciences),字节大小:168056.24297844546,样本数量:337 - 划分名称:其他(other),字节大小:177032.54082299152,样本数量:355 下载大小:813056 数据集总大小:1526968.0 配置列表: - 配置名称:基准(benchmark),数据文件: - 划分:测试集(test),路径:benchmark/test-* - 配置名称:训练(train),数据文件: - 划分:全量集(full),路径:train/full-* - 划分:理工科(STEM),路径:train/STEM-* - 划分:人文学科(humanities),路径:train/humanities-* - 划分:社会科学(social_sciences),路径:train/social_sciences-* - 划分:其他(other),路径:train/other-*



