anon-betterbench/betterbench-inference-logs

该数据集包含多个特征字段，包括开始时间、系统提示、用户输入、模型输出、模型名称、温度和推理时间。数据集主要用于记录和存储与模型交互的数据，可能用于训练或评估自然语言处理模型。数据集包含一个训练分割，共有1,806,398个样本，总大小为15,018,193,260字节。

This dataset is used to record and analyze interaction data of natural language processing models, including model outputs at different temperature parameters, user inputs, system prompts, and inference times. The dataset is primarily used for training models, containing 1806398 samples, with a dataset size of 15018193260 bytes.