Self-GRIT/mini-wikitext-preprocessed-PI_KFI-claude-3-5-sonnet-20240620

该数据集包含文本、消息和输出检索等字段。文本字段为字符串类型，消息字段包含内容和角色两个子字段，均为字符串类型。输出检索字段包含更新后的段落、错误代码和结果列表，结果列表中包含解释、理想知识、关键事实信息、前缀重构、前缀意图和验证等子字段。数据集包含一个训练集，共有10个样本，总大小为151868字节。数据集的下载大小为73579字节。

The dataset includes fields such as text, messages, and outputs_retrievals. The text field is of string type, the messages field contains two subfields: content and role, both of which are string types. The outputs_retrievals field includes updated passages, error codes, and a list of results, which contains subfields such as Explanation, Ideal Knowledge, Key Factual Information, Prefix reformulated, Prefix-intent, and Verification. The dataset contains a training set with 10 samples, totaling 151868 bytes. The download size of the dataset is 73579 bytes.

原始信息汇总

数据集概述

数据结构

• text:
  • 类型: string
• messages:
  • 类型: list
  • 子字段:
    • content:
      • 类型: string
    • role:
      • 类型: string
• outputs_retrievals:
  • 类型: struct
  • 子字段:
    • Updated_Passage:
      • 类型: string
    • error:
      • 类型: int64
    • results:
      • 类型: list
      • 子字段:
        • Explanation:
          • 类型: string
        • Ideal Knowledge:
          • 类型: string
        • Key Factual Information:
          • 类型: string
        • Prefix reformulated:
          • 类型: string
        • Prefix-intent:
          • 类型: string
        • Verification:
          • 类型: string

数据集划分

• train:
  • 字节数: 151868
  • 样本数: 10

数据集大小

• 下载大小: 73579 字节
• 数据集大小: 151868 字节

配置

• default:
  • 数据文件路径: data/train-*