GenuineWWD/SCS_data
收藏Hugging Face2025-11-14 更新2025-11-15 收录
下载链接：
https://hf-mirror.com/datasets/GenuineWWD/SCS_data
下载链接
链接失效反馈官方服务：
资源简介：
SCS数据集是一个用于多模态大型语言模型结果奖励强化学习训练的数据集，它通过自我一致性采样方法改进学习过程，在多选推理任务中提高模型的准确性和泛化能力。
The SCS dataset is a dataset for training outcome-reward reinforcement learning for multimodal large language models, which improves the learning process through self-consistency sampling method, enhancing the models accuracy and generalization ability in multiple-choice reasoning tasks.
提供机构：
GenuineWWD
