COLIEE

该数据集名为COLIEE，旨在评估基于日本律师资格考试的法律推理任务，特别关注蕴含推理。它包含了假设-答案对，这些对用于在GPT-3模型上进行零样本学习和少样本学习实验。该数据集规模较大，包含了不同年份的多个测试集，主要任务为法律推理和蕴含验证。The dataset is named COLIEE, which is designed to evaluate legal reasoning tasks based on the Japanese Bar Examination, with a special focus on entailment reasoning. It contains hypothesis-answer pairs, which are utilized for zero-shot and few-shot learning experiments on the GPT-3 model. This is a large-scale dataset incorporating multiple test sets from different years, whose primary tasks are legal reasoning and entailment verification.

搜集汇总

背景与挑战

背景概述

COLIEE数据集是一个专注于法律推理任务的数据集，特别针对日本律师资格考试中的蕴含推理进行设计。它包含假设-答案对，用于支持GPT-3模型的零样本和少样本学习实验，规模较大且涵盖多个年份的测试集。该数据集的核心目标是评估法律推理和蕴含验证能力。