rubricreward/PolyGuardMix-tgt_prompt_tgt_thinking
收藏官方服务：
资源简介：
该数据集包含文本输入和对应的模型输出，包括推理和响应。数据集中的字段包括标识符、语言、来源、实际分数、输入文本、提示令牌长度以及模型GPT-oss-120b的推理和响应。训练集包含约287万条示例。
The dataset includes text inputs and corresponding model outputs, consisting of reasoning and responses. The fields in the dataset include identifier, language, source, actual score, input text, prompt token length, and the reasoning and response from the GPT-oss-120b model. The training set contains approximately 2.87 million examples.
提供机构：
rubricreward
