SIRAC
收藏资源简介:
SIRAC是一个半结构化的IRAC(问题、规则、应用和结论)语料库,专门用于法律场景。该数据集包含了马来西亚合同法和澳大利亚社会法中关于依赖儿童的场景,每个场景都以半结构化格式进行了完整的IRAC分析,便于机器和法律专业人士解读和理解。
SIRAC is a semi-structured IRAC (Issue, Rule, Application, and Conclusion) corpus specifically designed for legal scenarios. This dataset encompasses scenarios related to dependent children under Malaysian contract law and Australian social law. Each scenario is thoroughly analyzed in a semi-structured IRAC format, facilitating interpretation and comprehension by both machines and legal professionals.
数据集概述
数据集名称
- SIRAC: 半结构化IRAC(Issue, Rule, Application and Conclusion)法律场景语料库
数据集内容
- 包含马来西亚合同法和澳大利亚依赖儿童社会法相关的法律场景。
- 每个场景都以半结构化格式标注了完整的IRAC分析,便于机器和法律专业人士理解和解释。
数据集结构
- 数据集文件结构示例: json { "id": "s01", "type": "Contract Law", "scenario": "...", "issue": "...", "dceompose question": "...", "analysis": "...", "conclusion": "yes, there is a valid contract between Alan and Cate" }
数据集统计
- SIRAC_ASA: 30个场景,1个问题,3条规则,平均推理长度4.8。
- SIRAC_CAM: 20个场景,20个问题,55条规则,平均推理长度9.3。
- SIRAC: 总计50个场景,21个问题,58条规则,平均推理长度7.05。
数据集使用
- 数据集可通过以下命令加载: python pip install git+https://github.com/christinakang/SIRAC.git python ./script/readScenario.py
数据集引用
-
引用格式:
@inproceedings{ anonymous2023can, title={Can Chat{GPT} Perform Reasoning Using the {IRAC} Method in Analyzing Legal Scenarios Like a Lawyer?}, author={Anonymous}, booktitle={The 2023 Conference on Empirical Methods in Natural Language Processing}, year={2023}, url={https://openreview.net/forum?id=7okuG5JhaM} }




