tikeape/localbench
收藏官方服务:
资源简介:
这是一个为本地模型扩展的简单基准测试数据集。虽然其中的问题可能并非100%正确,但由于创建这些问题的模型较为智能,该数据集仍可作为小型模型的有效基准测试工具。
Simple Bench but expanded a for local models. Questions may not be 100% correct as they are for local ai models but considering the model that created the question is smarter it should still be a effective benchmark for tiny models.
提供机构:
tikeape


