OALL/details_maldv__Qwenstein2.5-32B-Instruct

该数据集是在评估模型 [maldv/Qwenstein2.5-32B-Instruct](https://huggingface.co/maldv/Qwenstein2.5-32B-Instruct) 的过程中自动创建的。数据集由 136 个配置组成，每个配置对应一个评估任务。数据集由一次运行创建，每次运行在各个配置中都有一个特定的分割，分割的名称使用运行的时戳。train 分割始终指向最新的结果。还有一个名为 results 的附加配置，用于存储运行的所有汇总结果。文件还包括使用 Python 和 datasets 库加载数据集细节的示例，以及特定运行的最新结果详情。

Dataset automatically created during the evaluation run of model [maldv/Qwenstein2.5-32B-Instruct](https://huggingface.co/maldv/Qwenstein2.5-32B-Instruct). The dataset consists of 136 configurations, each corresponding to a different evaluated task. It is noted that the dataset was created from a single run, and each run is represented as a specific split within each configuration, named by the timestamp of the run. The train split always points to the latest results. There is an additional configuration called results which stores all the aggregated results of the run. The file also includes examples of how to load the dataset details using Python and the datasets library, as well as details of the latest results from a specific run.