Task 10 Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://competitions.codalab.org/competitions/17326
下载链接
链接失效反馈官方服务：
资源简介：
该数据集专为任务10竞赛设计，专注于基于翻译相似性预测的二元分类。对于任务10的预测，通过使用0.5的阈值将其转换为二元类别，达到了0.47的F1分数。数据集来源于结合模型或单独模型预测，其任务是进行情感检测的二元分类。
This dataset is specifically designed for the Task 10 competition, focusing on binary classification based on translation similarity prediction. For Task 10 predictions, a threshold of 0.5 was used to convert the outputs into binary categories, achieving an F1 score of 0.47. The dataset is sourced from predictions generated by either combined or standalone models, with its target task being binary classification for sentiment detection.
提供机构：
CodaLab Competitions
