RLHF-And-Friends/Human-vs-CoPPO
收藏Hugging Face2025-03-20 更新2025-04-26 收录
下载链接:
https://hf-mirror.com/datasets/RLHF-And-Friends/Human-vs-CoPPO
下载链接
链接失效反馈官方服务:
资源简介:
该数据集包含来自RLHF-And-Friends/tldr-sft测试分割的人类完成数据和RLHF-And-Friends/TLDR-Mistral-7B-SmallSFT-CoPPO模型的完成数据。prompt列包含同时提供给人类和模型的提示。使用的模型是TLDR-Mistral-7B-SmallSFT-CoPPO。基于gpt-4o-mini观点,该模型相对于人类的胜率为0.74。
This dataset contains human completions from the RLHF-And-Friends/tldr-sft test split and completions from the RLHF-And-Friends/TLDR-Mistral-7B-SmallSFT-CoPPO model. The prompt column includes the prompts given to both humans and the model. The model used is TLDR-Mistral-7B-SmallSFT-CoPPO. The winrate over humans based on the gpt-4o-mini opinion is 0.74.
提供机构:
RLHF-And-Friends



