SecureFinAI-Lab/Regulations_Link_Retrieval

该问题集旨在评估语言模型（LLM）检索特定法律规定的确切链接的能力。该数据集用于COLING 2025的监管挑战中的链接检索任务。目标是评估LLM在复杂法律数据库中导航以引用正确文档的能力。文本提到了需要此类引用的文档类型以及此能力对于法律研究和合规的重要性。它还提供了关于数据集的统计数据，详细说明了与不同监管类别相关的问题数量及其来源。使用的评估指标是准确度，数据集根据CDLA-Permissive-2.0许可进行授权，允许免费使用、修改和分发。

This question set is designed to assess the ability of Language Models (LLMs) to retrieve exact links to specific legal regulations. The dataset is intended for the link retrieval task at the Regulations Challenge @ COLING 2025. The objective is to evaluate how well LLMs can navigate through complex legal databases to reference the correct documents. The text mentions the types of documents that require such references and the importance of this ability for legal research and compliance. It also provides statistics on the dataset, detailing the count of questions related to different categories of regulations and their sources. The evaluation metric used is accuracy, and the dataset is licensed under CDLA-Permissive-2.0, which allows free use, modification, and redistribution.