zjhhhh/game_stage1_zjhhhh__qwen2.5_3B_Instruct_multi_stage1_seed_555134_eta_1e4_step_301
收藏数据链接:
官方服务:
资源简介:
该数据集包含了多个文本字段和评分字段,用于记录不同模型的响应和人类评分者的评分。具体包括提示(prompt)、要求(requirements)、基础响应(base_response)、模型响应(model_response)以及评分者的一致性评分(judge)等。数据集分为训练集(train),包含200个示例。
The dataset contains multiple text fields and rating fields, used to record the responses of different models and the ratings from human raters. It includes prompt, requirements, base responses (base_response), model responses (model_response), and consensus ratings from raters (judge). The dataset is split into a training set (train) with 200 examples.
提供机构:
zjhhhh


