bootcamp-labs/CSAR-2
收藏Hugging Face2025-10-20 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/bootcamp-labs/CSAR-2
下载链接
链接失效反馈官方服务：
资源简介：
CSAR²（Counter-Strike属性识别与推理）是一个全面的评估基准，用于测试模型对Counter-Strike 2在文本知识、视觉分析和音频事件分类方面的理解能力。 文本知识部分包含关于Counter-Strike 2的多个选择题，涉及地图、经济、手榴弹、武器和回合等类别。 模型需要回答这些问题来测试它们的CS2知识。
CSAR² (Counter-Strike Attributes Recognition & Reasoning) is a comprehensive evaluation benchmark designed to test model understanding of Counter-Strike 2 across textual knowledge, visual analysis capabilities, and audio event classification. The textual knowledge section includes multiple-choice questions about Counter-Strike 2, covering categories such as Map, Economy, Grenade, Weapon, and Round. Models are required to answer these questions to test their CS2 knowledge.
提供机构：
bootcamp-labs
