mmosbach/lm-eval-details

该数据集是在评估模型meta-llama/Llama-3.2-3B的过程中自动创建的。数据集由0个配置组成，每个配置对应一个评估任务。数据集从2次运行中创建，每次运行作为一个特定的分割，分割名称使用运行的时间戳。train分割始终指向最新的结果。此外，还有一个额外的配置results存储所有运行的聚合结果。数据集用于存储模型在任务（如gsm8k）上的评估细节和结果。

The dataset was automatically created during the evaluation run of the model meta-llama/Llama-3.2-3B. The dataset is composed of 0 configurations, each corresponding to one of the evaluated tasks. The dataset has been created from 2 runs, with each run being a specific split named by the timestamp of the run. The train split always points to the latest results. Additionally, there is an additional configuration results that stores all the aggregated results of the runs. The dataset is used to store evaluation details and results of the models performance on tasks such as gsm8k.