KenithZ/ProverQA-En2Zh

ProverQA是一个高质量的一阶逻辑(FOL)推理数据集，由ProverGen框架创建，该框架将大型语言模型(LLMs)的生成能力与符号证明者的严谨性和精确性相结合。这个数据集旨在评估和提高语言模型的逻辑推理能力，特别是在链式思维(CoT)环境中。数据集包含测试集和训练集，测试集包含1,500个实例，分为简单、中等和困难三个难度级别；训练集包含5,000个英文实例，用于微调实验。数据格式为JSON，包含id、options、answer、question、reasoning、context、nl2fol、conclusion_fol和word_mapping等字段，详细描述了每个实例的逻辑推理过程。

ProverQA is a high-quality First-Order Logic (FOL) reasoning dataset created by the ProverGen framework, which combines the generative capabilities of large language models (LLMs) with the rigor and precision of symbolic provers. This dataset is designed to evaluate and enhance the logical reasoning capabilities of language models, particularly in chain-of-thought (CoT) settings. The dataset includes a test set and a training set. The test set contains 1,500 instances divided into three difficulty levels: easy, medium, and hard. The training set contains 5,000 English instances for fine-tuning experiments. The data format is JSON, including fields such as id, options, answer, question, reasoning, context, nl2fol, conclusion_fol, and word_mapping, which detail the logical reasoning process for each instance.