英文

Pygmalion 7B

一个对话 LLaMA 细调模型

模型细节

Pygmalion 7B 是基于 Meta 的 LLaMA-7B 的对话模型。

这是版本 1。它使用来自 Pygmalion-6B-v8-pt4 的子数据进行了细调,对于熟悉该项目的人来说。

应用逻辑异或

此模型的 XOR 文件已经预先应用,直接使用即可。从 PygmalionAI 的发布版本 https://huggingface.co/PygmalionAI/pygmalion-7b 转换而来的 XOR 权重。

提示

模型是基于通常的 Pygmalion 个人信息 + 聊天格式进行训练的,因此任何通常的用户界面应该已经正确处理了所有内容。如果您直接使用模型,则应按以下格式进行操作:

[CHARACTER]'s Persona: [A few sentences about the character you want the model to play]
<START>
[DIALOGUE HISTORY]
You: [User's input message here]
[CHARACTER]:

在其中,[CHARACTER] 可以猜到,是您希望模型扮演的角色的名称,<START> 应该直接使用作为分隔符令牌,将个人信息和情景数据与对话分隔开来,[DIALOGUE HISTORY] 是聊天历史的滑动窗口,让模型可以从中获取对话背景。以下是一个具体的示例:

Assistant's Persona: Assistant is a highly intelligent language model trained to comply with user requests.
<START>
Assistant: Hello! How may I help you today?
You: What is Zork?
Assistant:

这将生成类似于以下内容:

 Zork is an interactive fiction computer game created in the 1970s by Infocom, Inc., which was later acquired by Activision Blizzard. It is widely considered one of the most influential games ever made and has been credited with popularizing text-based adventure games. The original version of Zork was written in the programming language MACRO-10, but it was ported to many other platforms over the years."

当模型判断响应已经完成时,将自动发出文本结束令牌(</s>)。

局限和偏见

此模型的预期用途是用于娱乐目的的虚构对话。任何其他类型的用途都超出了范围。

因此,它并没有被细调为安全和无害的:基础模型和此细调模型已经在包含粗话和淫秽或其他令人不悦的文本的数据上进行了训练。即使提示本身不包含任何明确的冒犯性内容,它可能产生社会上不可接受或不希望的文本。输出结果可能经常是事实错误或具有误导性的。