cmu-lti/osim-post-training
收藏资源简介:
SOUL是一个用于人类行为模拟的数据套件,应用于论文《通过口头反馈强化人类行为模拟》(DITTO),覆盖对话、社交模拟、社交认知、角色扮演和以人为中心的评估。该数据集是CMU-LTI组织对原始sunweiwei/Soul数据集的镜像,包含训练和测试数据,以parquet格式组织,每个任务都有独立的配置。数据用于支持ODYSSIM版本的后续训练,并包括多个任务如alignx、behaviorchain、coser等,旨在模拟人类行为以改进AI模型的社会智能和角色扮演能力。
SOUL is the data suite for human behavior simulation used in Reinforcing Human Behavior Simulation via Verbal Feedback (DITTO), spanning conversation, social simulation, social cognition, role-play, and human-centric evaluation. This CMU-LTI mirror hosts the post-training data used for ODYSSIM releases, mirroring the original sunweiwei/Soul dataset layout. It includes training and test data in parquet format, organized into multiple task configurations such as alignx, behaviorchain, coser, etc., designed to simulate human behavior for enhancing AI models social intelligence and role-play capabilities.




