hh-rlhf-ja
收藏arXiv2025-09-30 收录
下载链接：
https://huggingface.co/datasets/llm-jp/hh-rlhf-12k-ja
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是从hh-rlhf中抽取的样本，并使用DeepL翻译成了日语。它包含了12,000个实例，这些实例是为了公开可用而抽取的。该数据集的任务是来自人类反馈的强化学习。
This dataset comprises samples extracted from hh-rlhf, which were translated into Japanese using DeepL. It contains 12,000 instances curated for public accessibility. The target task of this dataset is reinforcement learning from human feedback (RLHF).
提供机构：
LLM-jp
