dogtooth/uf_Llama-3.1-Tulu-3-8B-SFT_8

该数据集包含用于训练和评估模型的多轮对话数据。每个样本包括一个提示（prompt）、一组消息（messages，每条消息包含内容和角色信息）、模型生成的完成内容（model_completion）以及参考完成内容（reference_completion）。数据集分为一个训练集，包含489,080个样本，总大小为1,808,763,360字节。

This dataset contains multi-turn dialogue data for training and evaluating models. Each sample includes a prompt, a set of messages (each containing content and role information), model-generated completions (model_completion), and reference completions (reference_completion). The dataset is divided into a training set containing 489,080 samples with a total size of 1,808,763,360 bytes.