bz-arc13/fc_rm_eval_pair_v1

该数据集包含多个特征，如functions、conversation、chosen和rejected，每个特征都有其特定的数据类型和结构。数据集分为三个部分：en_glaive、en_tool_learning和zh_hard，每个部分都有对应的字节大小和示例数量。此外，还提供了数据集的下载大小和总大小信息。

The dataset contains multiple features such as functions, conversation, chosen, and rejected, each with its specific data type and structure. The dataset is divided into three parts: en_glaive, en_tool_learning, and zh_hard, each with corresponding byte sizes and example counts. Additionally, the download size and total size of the dataset are provided.

原始信息汇总

数据集概述

数据集结构

• features:
  • functions:
    • dtype: string
  • conversation:
    • list:
      • content:
        • dtype: string
      • function_call:
        • struct:
          • arguments:
            • dtype: string
          • name:
            • dtype: string
      • name:
        • dtype: string
      • role:
        • dtype: string
  • chosen:
    • struct:
      • content:
        • dtype: string
      • function_call:
        • struct:
          • arguments:
            • dtype: string
          • name:
            • dtype: string
      • name:
        • dtype: null
      • role:
        • dtype: string
  • rejected:
    • struct:
      • content:
        • dtype: string
      • function_call:
        • struct:
          • arguments:
            • dtype: string
          • name:
            • dtype: string
      • name:
        • dtype: null
      • role:
        • dtype: string

数据集分割

• splits:
  • en_glaive:
    • num_bytes: 148894
    • num_examples: 125
  • en_tool_learning:
    • num_bytes: 1254547
    • num_examples: 259
  • zh_hard:
    • num_bytes: 2318703
    • num_examples: 234

数据集大小

• download_size: 381133
• dataset_size: 3722144

配置

• configs:
  • config_name: default
    • data_files:
      • split: en_glaive
        • path: data/en_glaive-*
      • split: en_tool_learning
        • path: data/en_tool_learning-*
      • split: zh_hard
        • path: data/zh_hard-*