violetxi/judge_calibration_int_qwen8b
收藏资源简介:
该数据集是一个用于训练和评估人工智能模型的数据集,包含1076个训练示例,总大小为9,812,118字节。数据集中每个样本包括多个特征:数据来源(data_source)、提示(prompt,包含角色和内容)、能力类型(ability)、额外信息(extra_info,如索引、拆分、解决方案、问题和答案)、响应(response)、是否包含思考过程(has_thinking)和评分(score)。数据集支持训练拆分,并采用默认配置,数据文件位于data/train-*路径下。
This dataset is designed for training and evaluating AI models, containing 1,076 training examples with a total size of 9,812,118 bytes. Each sample in the dataset includes multiple features: data source, prompt (with role and content), ability type, extra information (such as index, split, solution, question, and answer), response, whether it includes thinking process (has_thinking), and score. The dataset supports a training split and uses a default configuration with data files located at the data/train-* path.




