PAug/ProofNetVerif
收藏Hugging Face2025-02-06 更新2025-02-15 收录
下载链接:
https://hf-mirror.com/datasets/PAug/ProofNetVerif
下载链接
链接失效反馈官方服务:
资源简介:
ProofNetVerif是一个评估陈述自动形式化的基准数据集,用于评估基于参考和无需参考的度量。数据集包含id、自然语言陈述、Lean4源代码头部、Lean4形式化、Lean4预测和正确性等字段。数据集分为valid和test两个部分,分别包含2300和1452个示例。
ProofNetVerif is a benchmark for evaluating statement autoformalization, assessing both reference-based and reference-free metrics. The dataset includes fields such as id, natural language statements, Lean4 source header, Lean4 formalization, Lean4 prediction, and correctness. It is split into two parts, valid and test, containing 2300 and 1452 examples respectively.
提供机构:
PAug



