OlympicArena

该数据集名为“奥林匹克竞技场”，是一个极具挑战性的数据集，它用于评估模型在弱到强学习框架下的推理能力。在向前看的实验设置中，该数据集被用于让Llama3-8b-instruct模型指导Llama3-70b模型，以完成复杂的推理任务。

This dataset, named "Olympic Arena", is a highly challenging benchmark dataset designed to evaluate models' reasoning capabilities under the weak-to-strong learning framework. In the look-ahead experimental setting, this dataset is utilized to enable the Llama3-8b-instruct model to guide the Llama3-70b model in completing complex reasoning tasks.