five

flaitenberger/proofwriter_gold_formalized

收藏
Hugging Face2026-04-01 更新2026-04-12 收录
下载链接:
https://hf-mirror.com/datasets/flaitenberger/proofwriter_gold_formalized
下载链接
链接失效反馈
官方服务:
资源简介:
--- dataset_info: features: - name: dataset dtype: string - name: subset dtype: 'null' - name: split dtype: string - name: index dtype: int64 - name: question dtype: string - name: formalized_prompt dtype: string - name: formalized_constants dtype: string - name: formalized_predicates dtype: string - name: formalized_premises dtype: string - name: natural_language_premises dtype: string - name: formalized_conclusion dtype: string - name: formalization_syntax_ok dtype: bool - name: formalization_format_ok dtype: bool - name: formalization_errors dtype: string - name: formalization_problematic dtype: bool - name: formalization_warnings dtype: string - name: formalization_turns_used dtype: int64 - name: formalization_self_judge_complete dtype: bool - name: formalization_self_judge_missing dtype: string - name: formalization_raw_output dtype: string - name: ground_truth dtype: string - name: reference_answers list: string - name: correctness_scorer dtype: string - name: value_normalizer dtype: string - name: source_row dtype: string splits: - name: train num_bytes: 2703780685 num_examples: 585552 download_size: 247401571 dataset_size: 2703780685 configs: - config_name: default data_files: - split: train path: data/train-* ---

数据集信息: 特征列表: - 字段名:dataset,数据类型:字符串 - 字段名:subset,数据类型:空类型(null) - 字段名:split,数据类型:字符串 - 字段名:index,数据类型:64位整型(int64) - 字段名:question,数据类型:字符串 - 字段名:formalized_prompt,数据类型:字符串,即形式化提示(formalized prompt) - 字段名:formalized_constants,数据类型:字符串,即形式化常量(formalized constants) - 字段名:formalized_predicates,数据类型:字符串,即形式化谓词(formalized predicates) - 字段名:formalized_premises,数据类型:字符串,即形式化前提(formalized premises) - 字段名:natural_language_premises,数据类型:字符串,即自然语言前提(natural language premises) - 字段名:formalized_conclusion,数据类型:字符串,即形式化结论(formalized conclusion) - 字段名:formalization_syntax_ok,数据类型:布尔型(bool),即形式化语法合规性 - 字段名:formalization_format_ok,数据类型:布尔型,即形式化格式合规性 - 字段名:formalization_errors,数据类型:字符串,即形式化错误信息 - 字段名:formalization_problematic,数据类型:布尔型,即形式化结果存在问题 - 字段名:formalization_warnings,数据类型:字符串,即形式化警告信息 - 字段名:formalization_turns_used,数据类型:64位整型,即使用的形式化交互轮次 - 字段名:formalization_self_judge_complete,数据类型:布尔型,即形式化自我判断完整性 - 字段名:formalization_self_judge_missing,数据类型:字符串,即形式化自我判断的缺失项 - 字段名:formalization_raw_output,数据类型:字符串,即形式化原始输出 - 字段名:ground_truth,数据类型:字符串,即标准答案(ground truth) - 字段名:reference_answers,数据类型:字符串列表(list: string),即参考答案列表 - 字段名:correctness_scorer,数据类型:字符串,即正确性评分器 - 字段名:value_normalizer,数据类型:字符串,即值归一化器 - 字段名:source_row,数据类型:字符串,即源数据行 数据集划分: - 划分名称:训练集(train),字节大小:2703780685,样本数量:585552 下载总大小:247401571 数据集总大小:2703780685 数据集配置: - 配置名称:默认配置(default),数据文件: - 划分:训练集,文件路径:data/train-*
提供机构:
flaitenberger
二维码
社区交流群
二维码
科研交流群
商业服务