遇见数据集

unlearning-cleanslate/eval-qwen3-8b-simnpo-gentle-bm25-10b

收藏
Hugging Face2026-04-29 更新2026-05-03 收录
官方服务:

资源简介:

该数据集包含文本记忆化评估结果,特征包括文本长度、窗口数量、记忆化窗口、记忆化分数、覆盖率、概率分数,以及详细的窗口信息如索引、目标、种子、对数概率等,用于分析语言模型的记忆化行为。

This dataset contains evaluation results for text memorization, including features such as text length, number of windows, memorized windows, memorized fraction, coverage, probability scores, and detailed window information like indices, targets, seeds, log probabilities, etc. It is used for analyzing memorization behavior of language models.

提供机构:
unlearning-cleanslate
二维码
社区交流群
二维码
科研交流群
商业服务