bz-arc13/preference_dataset_mixture2_and_safe_pku

该数据集包含三个主要特征：rejected、chosen和conversation。每个特征都包含content和role两个子特征，数据类型均为字符串。数据集分为一个训练集，包含554,903个样本，总大小为1,371,957,337字节。下载大小为808,672,599字节。

The dataset contains three main features: rejected, chosen, and conversation. Each feature includes two sub-features: content and role, both of which are of string type. The dataset is divided into one training set, containing 554,903 samples, with a total size of 1,371,957,337 bytes. The download size is 808,672,599 bytes.

原始信息汇总

数据集概述

数据集信息

特征

• rejected
  • content: 字符串类型
  • role: 字符串类型
• chosen
  • content: 字符串类型
  • role: 字符串类型
• conversation
  • content: 字符串类型
  • role: 字符串类型

数据分割

• train
  • num_bytes: 1371957337
  • num_examples: 554903

数据集大小

• download_size: 808672599
• dataset_size: 1371957337

配置

• config_name: default
  • data_files
    • split: train
    • path: data/train-*