rubricreward/PolyGuardMix-tgt_prompt_en_thinking-filtered_correct

该数据集包含了文本输入和对应的模型输出，具体特征包括文本id、语言类型、数据来源、实际得分、文本输入、提示符token长度以及GPT-oss-120b模型生成的推理和响应。数据集分为训练集，其大小为约11.1GB，共有2618235个示例。

This dataset includes text inputs and corresponding model outputs, with features such as text id, language type, data source, actual score, text input, prompt token length, and reasoning and responses generated by the GPT-oss-120b model. The dataset is split into a training set, which is approximately 11.1GB in size and contains 2,618,235 examples.