ArabLegalEval
收藏资源简介:
ArabLegalEval数据集包含三个配置:ArLegalBench、MCQs和QA。ArLegalBench配置包含法律相关的问题和答案,MCQs配置包含多项选择题,QA配置包含问答对。每个配置都有不同的特征和分割,例如问题、答案、上下文、选项等。数据集用于法律领域的自然语言处理任务,如问答和选择题的生成与评估。
The ArabLegalEval dataset comprises three configurations: ArLegalBench, MCQs, and QA. The ArLegalBench configuration encompasses legal-related questions and answers, the MCQs configuration contains multiple-choice questions, and the QA configuration consists of question-answer pairs. Each configuration has distinct characteristics and dataset splits, such as questions, answers, contexts, options, and the like. This dataset is utilized for natural language processing tasks in the legal domain, including the generation and evaluation of question answering and multiple-choice question tasks.
ArabLegalEval 数据集概述
数据集配置
ArLegalBench
- 特征:
index: 字符串question_english: 字符串answer_english: 字符串contract_english: 字符串Question: 字符串Answer: 字符串Context: 字符串choices: 字符串序列choices_english: 字符串序列subject: 字符串task: 字符串
- 数据分割:
test: 19,243,593 字节,15,311 个样本train: 51,939 字节,28 个样本
- 下载大小: 3,389,747 字节
- 数据集大小: 19,295,532 字节
MCQs
- 特征:
Engine: 字符串Context: 字符串Question: 字符串Answer: 字符串Option 1: 字符串Option 2: 字符串Option 3: 字符串Option 4: 字符串Answer Key: 整型task: 字符串subject: 字符串choices: 字符串序列
- 数据分割:
train: 12,185,640.889 字节,9,291 个样本test: 3,046,738.111 字节,2,323 个样本
- 下载大小: 7,038,064 字节
- 数据集大小: 15,232,379.0 字节
QA
- 特征:
Question: 字符串Answer: 字符串Context: 字符串task: 字符串subject: 字符串
- 数据分割:
train: 81,006.038 字节,63 个样本test: 20,572.962 字节,16 个样本
- 下载大小: 56,179 字节
- 数据集大小: 101,579.0 字节
数据文件路径
ArLegalBench
train: ArLegalBench/train-*test: ArLegalBench/test-*
MCQs
train: MCQs/train-*test: MCQs/test-*
QA
train: QA/train-*test: QA/test-*




