YanAdjeNole/eppc-0shot

该数据集是在评估模型 deepseek-ai/DeepSeek-R1-Distill-Qwen-32B 时自动创建的。数据集由 7 个配置组成，每个配置对应一个评估任务。数据集由 80 次运行创建，每次运行在每个配置中都有一个特定的分割，分割名称使用运行的时间戳。train 分割始终指向最新的结果。还有一个 results 配置，用于存储所有运行的汇总结果。数据集还包括从特定运行加载结果的说明，例如使用 2025-11-14T08-02-59.514888 运行的最新结果。数据集的详细信息部分大部分不完整，有关于由谁策划、由谁资助、由谁共享、语言、许可证和其他重要信息的占位符。数据集的结构、创建理由、源数据、注释、偏差、风险、限制和建议也被指示为需要更多信息。

The dataset was automatically created during the evaluation run of the model deepseek-ai/DeepSeek-R1-Distill-Qwen-32B. It consists of 7 configurations, each corresponding to a specific evaluated task. The dataset was generated from 80 runs, with each run available as a separate split named by the runs timestamp. A train split always points to the latest results. There is also a results configuration storing aggregated results. The dataset includes details on how to load results from a specific run, providing an example using the latest results from run 2025-11-14T08-02-59.514888. The dataset details section is mostly incomplete, with placeholders for curated by, funded by, shared by, languages, license, and other important information. The dataset structure, creation rationale, source data, annotations, bias, risks, limitations, and recommendations are also indicated as needing more information.