NinaCalvi/ultra-rm-truthfulness-1000-GRM3B

该数据集包含多个特征字段，主要用于评估模型的回答质量。数据集中的每个样本包含source（来源）、instruction（指令）、completions（完成情况）等信息。completions字段下包含annotations（注释）、critique（批评）、custom_system_prompt（自定义系统提示）等子字段。annotations字段进一步细分为helpfulness（有用性）、honesty（诚实性）、instruction_following（指令遵循）和truthfulness（真实性）等子字段，每个子字段包含Rating（评分）、Rationale（理由）等详细信息。数据集还包含correct_answers（正确答案）、incorrect_answers（错误答案）、split（分割）等字段。数据集包含一个train分割，包含2000个样本，文件大小为51029806字节。

This dataset contains multiple feature fields, primarily used to evaluate the quality of model responses. Each sample in the dataset includes information such as source, instruction, completions, etc. The completions field contains subfields such as annotations, critique, custom_system_prompt, etc. The annotations field is further divided into subfields like helpfulness, honesty, instruction_following, and truthfulness, each containing details such as Rating and Rationale. The dataset also includes fields like correct_answers, incorrect_answers, split, etc. The dataset contains a train split with 2000 samples and a file size of 51029806 bytes.