ebayes/uhura-truthfulqa

Uhura-TruthfulQA是一个多语言数据集，包含Amharic、Hausa、Northern Sotho、Swahili和Yoruba五种语言。该数据集主要用于多项选择题的问答任务，数据量在1K到10K之间。数据集的结构包括生成和多项选择两种任务类型，每种语言都有对应的训练和测试数据文件。

Uhura-TruthfulQA is a multilingual dataset that includes five languages: Amharic, Hausa, Northern Sotho, Swahili, and Yoruba. The dataset is primarily used for multiple-choice question-answering tasks, with a data size ranging between 1K and 10K. The dataset structure includes both generation and multiple-choice task types, with corresponding training and test data files for each language.

原始信息汇总

数据集概述

数据集名称

• 名称: Uhura-TruthfulQA

许可证

• 许可证: cc-by-nc-4.0

语言

• 支持语言:
  • Amharic
  • Hausa
  • Northern Sotho (Sepedi)
  • Swahili
  • Yoruba
  • Zulu

大小分类

• 数据集大小: 1K<n<10K

多语言性

• 多语言支持: 是

标签

• 标签:
  • uhura
  • truthfulqa

任务类别

• 任务类别: 多项选择

任务ID

• 任务ID: 多项选择-qa

配置

• 配置名称: am_generation, am_multiple_choice, ha_generation, ha_multiple_choice, sw_generation, sw_multiple_choice, yo_generation, yo_multiple_choice
• 数据文件:
  • split: train
  • path:
    • truthfulqa_am_gen.json
    • truthfulqa_am_mc1.json
    • truthfulqa_ha_gen.json
    • truthfulqa_ha_mc1.json
    • truthfulqa_sw_gen.json
    • truthfulqa_sw_mc1.json
    • truthfulqa_yo_gen.json
    • truthfulqa_yo_mc1.json