unlearning-cleanslate/not-exp-fsid-curated-qwen3-coder-next-target-100

该数据集包含多个配置（forget、forget_pool、retain、retain_pool），每个配置具有不同的特征和分割。forget配置包含请求ID、内容ID、内容标题、窗口索引、前缀、后缀、记忆分数和规则名称等特征。forget_pool配置包含内容ID、内容标题、内容创作者、内容年份、歌词、记忆分数等特征。retain配置包含文本和规则名称特征。retain_pool配置包含文本长度、窗口数量、记忆窗口、记忆分数、复制窗口、复制分数等特征。数据集主要用于与记忆、复制和评估相关的任务。

The dataset includes multiple configurations (forget, forget_pool, retain, retain_pool), each with distinct features and splits. The forget configuration features request ID, content ID, content title, window index, prefix, suffix, memorized fraction, and rule name. The forget_pool configuration includes content ID, content title, content creators, content year, lyrics, memorized fraction, etc. The retain configuration features text and rule name. The retain_pool configuration includes text length, number of windows, memorized windows, memorized fraction, reproduced windows, reproduced fraction, etc. The dataset is primarily used for tasks related to memorization, reproduction, and evaluation.