DanqingZ/huggingface_doc_qa_eval_5

该数据集包含多个字段，包括上下文（context）、问题（question）、答案（answer）、来源文档（source_doc）以及多个评分和评估字段（如groundedness_score、relevance_score等）。这些字段表明数据集可能用于问答系统的训练和评估，特别是关注答案的准确性、相关性和独立性。数据集包含一个训练分割，共有6个示例，总大小为21731字节。

This dataset is primarily used for evaluating and analyzing the performance of question answering systems. It includes multiple features such as context, question, answer, source document, and various scores and evaluations related to the answers. The dataset is divided into a training set, containing 6 samples, for training models. The download size of the dataset is 46984 bytes, and the actual size is 21731 bytes.

原始信息汇总

数据集概述

数据集信息

• 特征字段:
  • context: 类型为字符串
  • question: 类型为字符串
  • answer: 类型为字符串
  • source_doc: 类型为字符串
  • groundedness_score: 类型为int64
  • groundedness_eval: 类型为字符串
  • relevance_score: 类型为int64
  • relevance_eval: 类型为字符串
  • standalone_score: 类型为int64
  • standalone_eval: 类型为字符串

数据集划分

• train:
  • num_bytes: 21731
  • num_examples: 6

数据集大小

• download_size: 46984
• dataset_size: 21731

配置

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