OALL/details_meta-llama__Llama-3.2-3B-Instruct

这是一个在模型 meta-llama/Llama-3.2-3B-Instruct 的评估过程中自动创建的数据集。数据集包含136个配置，每个配置对应一个评估任务。数据集由一次运行生成，包括一个指向最新结果的 train 切片。还有一个额外的配置 results 存储了运行的所有汇总结果。README文件提供了一个使用Python和Hugging Face datasets库从运行中加载详细信息的示例。

This is a dataset automatically created during the evaluation run of the model meta-llama/Llama-3.2-3B-Instruct. The dataset consists of 136 configurations, each corresponding to an evaluated task. The dataset is derived from one run and includes a train split that points to the latest results. There is also an additional configuration named results that stores aggregated results from the run. The README provides an example of how to load details from a run using Python and the Hugging Face datasets library.