InsultedByMathematics/lr-5e-7_update_401_eval
收藏Hugging Face2025-01-30 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/InsultedByMathematics/lr-5e-7_update_401_eval
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含多个字符串响应和提示，以及与这些响应和提示相关的奖励和概率。具体包括：不同版本的响应和提示文本，选择、拒绝和中间选项的奖励，以及基础和微调LLAMA模型的标记序列和对应的对数概率。
The dataset contains multiple string responses and prompts, along with rewards and probabilities associated with these responses and prompts. It includes: different versions of response and prompt texts, rewards for chosen, rejected, and middle options, as well as token sequences and corresponding log probabilities for the base and finetuned LLAMA models.
提供机构：
InsultedByMathematics
