allura-org/fujin-filtered-noliterotica
收藏Hugging Face2025-03-12 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/allura-org/fujin-filtered-noliterotica
下载链接
链接失效反馈官方服务：
资源简介：
经过处理的富津数据集，移除了文学洛丽塔样本以突出富津的独特贡献。数据集经过了两个主要的过滤过程：lossfiltered 过滤掉了估计损失过高或过低的样本，以避免随机垃圾或人工智能生成的内容；delinked 过滤掉了包含过多URL的样本。
The processed fujin dataset with literotica samples removed to highlight the unique contributions of fujin. The dataset has undergone two main filtering processes: lossfiltered, which removes samples with excessively high or low estimated loss to avoid random jank or AI-generated content; delinked, which removes samples with an excessive number of URLs.
提供机构：
allura-org
