AIM-Harvard/race-injection-medqa
收藏资源简介：
--- dataset_info: features: - name: id dtype: string - name: sent1 dtype: string - name: sent2 dtype: string - name: ending0 dtype: string - name: ending1 dtype: string - name: ending2 dtype: string - name: ending3 dtype: string - name: label dtype: int64 splits: - name: train num_bytes: 9016446 num_examples: 10178 - name: asian num_bytes: 882359 num_examples: 960 - name: black num_bytes: 882359 num_examples: 960 - name: white num_bytes: 882359 num_examples: 960 - name: hispanic num_bytes: 885203 num_examples: 960 - name: vanilla num_bytes: 877619 num_examples: 960 download_size: 7623391 dataset_size: 13426345 configs: - config_name: default data_files: - split: train path: data/train-* - split: asian path: data/asian-* - split: black path: data/black-* - split: white path: data/white-* - split: hispanic path: data/hispanic-* - split: vanilla path: data/vanilla-* ---
The dataset includes multiple features such as id, sent1, sent2, ending0 to ending3, and label, primarily consisting of string and integer types. The dataset is divided into several subsets, including a training set and subsets for different races, each with corresponding byte size and number of examples. Additionally, the download size and actual size of the dataset are provided.
数据集概述
特征信息
- id: 数据类型为字符串。
- sent1: 数据类型为字符串。
- sent2: 数据类型为字符串。
- ending0: 数据类型为字符串。
- ending1: 数据类型为字符串。
- ending2: 数据类型为字符串。
- ending3: 数据类型为字符串。
- label: 数据类型为整数（int64）。
数据分割
- train: 包含10178个样本，占用9016446字节。
- asian: 包含960个样本，占用882359字节。
- black: 包含960个样本，占用882359字节。
- white: 包含960个样本，占用882359字节。
- hispanic: 包含960个样本，占用885203字节。
- vanilla: 包含960个样本，占用877619字节。
数据集大小
- 下载大小: 7623391字节。
- 数据集总大小: 13426345字节。
配置信息
- default配置包含以下数据文件路径：
- train:
data/train-* - asian:
data/asian-* - black:
data/black-* - white:
data/white-* - hispanic:
data/hispanic-* - vanilla:
data/vanilla-*
- train:
