prli/uspto-full_draft_chopped-rare-qwen-alpha0-90_perturb
收藏Hugging Face2026-04-27 更新2026-05-03 收录
下载链接:
https://hf-mirror.com/datasets/prli/uspto-full_draft_chopped-rare-qwen-alpha0-90_perturb
下载链接
链接失效反馈官方服务:
资源简介:
该数据集包含原始文本和通过多种文本扰动方法生成的变体文本,具体包括同义词替换、手指误触模拟、随机删除、字符大小写变化、空格扰动和下划线技巧等扰动方式。数据集仅提供验证集,包含4000个样本。
This dataset contains original texts and their variants generated by various text perturbation methods, including synonym substitution, butter fingers simulation, random deletion, character case changes, whitespace perturbation, and underscore tricks. The dataset only provides a validation set with 4000 examples.
提供机构:
prli



