unlearning-cleanslate/eval-02-qwen3-8b-simnpo-gentle-baseline-target-100-checkpoint-539

该数据集是一个用于评估文本记忆或语言模型性能的数据集，包含4663个训练示例，总大小约2.67GB。特征包括文本长度字符数、窗口数量、记忆窗口数、记忆分数、覆盖率、概率统计值（如最大、平均、最小p_z）、最佳窗口信息（如索引、概率、种子、目标文本）、评估模型、窗口大小、步长、评估阈值，以及详细窗口数据（如结束字符、索引、是否记忆、对数概率、目标令牌数、p_z值、种子、起始字符、目标文本、目标对数概率和目标排名）。此外，还包含内容ID、标题、创建者和年份等元数据。数据集仅提供训练分割，适用于NLP研究和模型评估任务。

This dataset is designed for evaluating text memorization or language model performance, containing 4663 training examples with a total size of approximately 2.67GB. Features include text length in characters, number of windows, memorized windows, memorized fraction, coverage, probability statistics (such as max, mean, min p_z), best window information (e.g., index, probability, seed, target text), evaluation model, window size, stride, evaluation threshold, and detailed window data (such as end character, index, is memorized, log probability, number of target tokens, p_z value, seed, start character, target text, target log probabilities, and target ranks). Additionally, it includes metadata like content ID, title, creators, and year. The dataset only provides a training split and is suitable for NLP research and model evaluation tasks.