chrisliu298/preference_dataset_mixture2_and_safe_pku30k_and_argilla_math_and_ultra_code_for_preference_model

该数据集与[weqweasdas/preference_dataset_mixture2_and_safe_pku30k_and_argilla_math_and_ultra_code_for_preference_model](https://huggingface.co/datasets/weqweasdas/preference_dataset_mixture2_and_safe_pku30k_and_argilla_math_and_ultra_code_for_preference_model)相同，但已按照标准格式处理。数据集包含chosen和rejected两个主要特征，每个特征包含content和role两个子特征，数据类型均为字符串。数据集包含一个训练集分割，包含606,493个例子，总大小为2,060,273,480字节。

This dataset is the same as [weqweasdas/preference_dataset_mixture2_and_safe_pku30k_and_argilla_math_and_ultra_code_for_preference_model](https://huggingface.co/datasets/weqweasdas/preference_dataset_mixture2_and_safe_pku30k_and_argilla_math_and_ultra_code_for_preference_model) but has been processed in the standard format. The dataset includes two main features: chosen and rejected, each containing two sub-features: content and role, with data types as string. The dataset contains one training split with 606,493 examples, totaling 2,060,273,480 bytes.

原始信息汇总

数据集概述

数据集信息

• 特征:

  • chosen:
    • content: 数据类型为字符串
    • role: 数据类型为字符串
  • rejected:
    • content: 数据类型为字符串
    • role: 数据类型为字符串

• 数据分割:

  • train:
    • num_bytes: 2060273480 字节
    • num_examples: 606493 个样本

• 下载大小: 1145772728 字节

• 数据集大小: 2060273480 字节

配置

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