Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Feature Suggestion: The chat interface should include "editable system messages, user responses, and AI responses." #855

Open
Wooden-Gear opened this issue Nov 26, 2024 · 2 comments

Comments

@Wooden-Gear
Copy link

功能建议:聊天界面可添加“可编辑的系统消息、用户回复、AI回复”
如下图:
image
图中的输入框都是可编辑的,这个功能有利于复杂提示词的输入,可以更好的规范AI的输出。

Feature suggestion: Add editable system messages, user replies, and AI replies to the chat interface.
The input boxes in the image are all editable. This feature facilitates the input of complex prompts and allows for better regulation of AI output.

@logancyang
Copy link
Owner

System message and user messages are editable already. Is this asking for editable AI response?

@Wooden-Gear
Copy link
Author

System message and user messages are editable already. Is this asking for editable AI response?

上图是某个软件的AI聊天框,可以随意添加上下文,并设置这段添加的上下文是系统消息,还是用户回复,还是AI回复。这样能够欺骗AI,让AI以为某一段文字是属于以上的三种类型中的哪种。

这样的好处是,在有些冗长的对话中,我如果选择了某段AI的回答,但又需要人工做一些修改之后,才能继续后续的聊天,那就可以实时修改,并删除其他不必要的聊天内容以减少上下文。之后可以在我修改的那段AI回复的基础上,继续聊天,让AI以为那段回复是他自己做出的。

这种操作的优点是:

  1. 减少上下文内容,防止AI抓不住重点。
  2. 人工修改更精确,让后续的聊天更明确。

The image shows an AI chatbox from a certain software. You can freely add context and specify whether the added context is a system message, a user reply, or an AI reply. This allows you to deceive the AI into believing a certain text belongs to one of these three types.

The benefit is that, in lengthy conversations, if I select an AI response but need to manually modify it before continuing the chat, I can modify it in real-time and delete unnecessary chat content to reduce the context. Then, I can continue the chat based on my modified AI response, making the AI believe it produced that response itself.

Advantages of this operation are:

  1. Reducing context content prevents the AI from missing the point.
  2. Manual modification is more precise, making subsequent conversations clearer.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants