julien31/soar_arc_train_5M
收藏Hugging Face2025-07-25 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/julien31/soar_arc_train_5M
下载链接
链接失效反馈官方服务：
资源简介：
SOAR-ARC模型数据集包含了大约500万个ARC解决方案。对于成功解决原始ARC任务 solution，通过代码进行去重以保证唯一性。对于通过后见标签化生成的新的合成任务的solution，则根据输出结果进行去重。这种方式确保了数据集的多样性和高质量，适用于进一步的研究和开发。
The SOAR-ARC dataset contains around 5 million ARC solutions. For solutions that successfully solve an original ARC task, entries are deduplicated by their code to ensure uniqueness. For solutions corresponding to new synthetic tasks generated via hindsight relabeling, deduplication is based on their output results. This approach ensures a diverse and high-quality dataset for further research and development.
提供机构：
julien31
