MisDrifter/1019_Qwen__Qwen2.5-3B-Instruct
收藏Hugging Face2025-10-20 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/MisDrifter/1019_Qwen__Qwen2.5-3B-Instruct
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了一系列对话场景，每个场景包括一个提示(prompt)、要求(requirements)和两个模型的响应(model_response_0, model_response_1)。每个响应后面跟着两个评判结果，一个是大多数评判者的意见(judge_0_0_majority, judge_1_0_majority)，另一个是评判的平均值(judge_0_0_mean, judge_1_0_mean)。数据集目前只有一个训练集分割，大小为96936字节，共有10个示例。
The dataset consists of a series of dialogue scenarios, each including a prompt, requirements, and responses from two models (model_response_0, model_response_1). Each response is followed by two judgment results, one is the opinion of the majority of judges (judge_0_0_majority, judge_1_0_majority), and the other is the average of the judgments (judge_0_0_mean, judge_1_0_mean). The dataset currently has only one training split, which is 96936 bytes in size and contains 10 examples.
提供机构：
MisDrifter
