isaacus/mleb-scalr

SCALR (MLEB版本)是一个用于评估信息检索模型在检索与复杂、推理密集型法律问题相关的法律条文方面的能力的测试数据集。数据集包括三个部分：default部分将问题与正确的法律条文对配对，每个对有一个分数为1；corpus部分包含法律条文，每个条文有一个唯一的标识符和标题（标题故意设置为空字符串以兼容mteb库）；queries部分包含问题。

SCALR (MLEB version) is a test dataset for evaluating the ability of information retrieval models to retrieve relevant legal holdings for complex, reasoning-intensive legal questions. The dataset includes three parts: the default part pairs questions with correct legal holdings, each pair having a score of 1; the corpus part contains legal holdings, each with a unique identifier and a title (which is intentionally set to an empty string for compatibility with the mteb library); the queries part contains questions.