SystemChat

This dataset by AbacusAI was crafted by Eric Hartford

This is a synthetic dataset, generated mainly with Mistral-Medium and [dolphin-2.7-mixtral-8x7b](https://huggingface.co/cognitivecomputations/dolphin-2.7-mixtral-8x7b)

The purpose of this dataset is to train the model to respect the System Prompt throughout the entire conversation, no matter how unconventional the system prompt might be.

This dataset is under continued development - my intent is to grow it to 100k conversations.

But, for now, it is good enough to start using.

#### 下载方法
:modelscope-code[]{type="sdk"}
:modelscope-code[]{type="git"}

本数据集由AbacusAI出品，由Eric Hartford精心打造。
该数据集为合成数据集，主要依托Mistral-Medium与[dolphin-2.7-mixtral-8x7b](https://huggingface.co/cognitivecomputations/dolphin-2.7-mixtral-8x7b)生成。
本数据集的核心用途为训练模型，使其能够在完整对话的全流程中严格遵循系统提示（System Prompt），无论该系统提示的设定何等非常规。
本数据集仍处于持续迭代开发阶段，制作者计划将其扩充至10万条对话的规模。
但就当前阶段而言，该数据集已足以投入实际使用。
#### 下载方法
:modelscope-code[]{type="sdk"}
:modelscope-code[]{type="git"}