yueliu1999/GuardReasonerTrain
收藏Hugging Face2025-02-02 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/yueliu1999/GuardReasonerTrain
下载链接
链接失效反馈官方服务：
资源简介：
GuardReasonerTrain是GuardReasoner的R-SFT训练数据，用于支持基于推理的大型语言模型的安全防护。该数据集包含四个分割：WildGuardTrainR、AegisTrainR、BeaverTailsTrainR和ToxicChatTrainR，并以JSON格式存储。它适用于文本分类任务，支持英语，数据规模在10万到100万条之间。
GuardReasonerTrain is the training data for R-SFT of GuardReasoner, which is used to support reasoning-based safeguards for large language models. The dataset includes four splits: WildGuardTrainR, AegisTrainR, BeaverTailsTrainR, and ToxicChatTrainR, stored in JSON format. It is suitable for text classification tasks, supports English, and has a size range of 100,000 to 1,000,000 entries.
提供机构：
yueliu1999
