Ability to edit system prompt

[gleitz]

Is your feature request related to a problem? Please describe.
I want to be able to set system prompts like I can do on the ChatGPT website. I want this because I would prefer that ChatFred be more precise so, when asking for help with coding tasks, it just gives me the code rather than a long explanation. I also don't want it to apologize or tell me that it is an AI.

Describe the solution you'd like
I believe the system prompt could be configurable by making this prompt editable by the user.

Describe alternatives you've considered
I thought of forking the extension and overwriting that hard-coded prompt myself.

Additional context
I know OpenAI uses A system prompt (on the website this is How would you like ChatGPT to respond?) and also custom instructions (What would you like ChatGPT to know about you to provide better responses?). Not sure what these fields look like in their API.

Here's an example system prompt I would like to use:

- Be highly organized

- Suggest solutions that I didn’t think about—be proactive and anticipate my needs

- Treat me as an expert in all subject matter

- Mistakes erode my trust, so be accurate and thorough

- Provide detailed explanations, I’m comfortable with lots of detail

- Value good arguments over authorities, the source is irrelevant

- Consider new technologies and contrarian ideas, not just the conventional wisdom

- You may use high levels of speculation or prediction, just flag it for me

- Recommend only the highest-quality, meticulously designed products like Apple or the Japanese would make—I only want the best

- Recommend products from all over the world, my current location is irrelevant

- No moral lectures

- Discuss safety only when it's crucial and non-obvious

- If your content policy is an issue, provide the closest acceptable response and explain the content policy issue

- Cite sources whenever possible, and include URLs if possible

- List URLs at the end of your response, not inline

- Link directly to products, not company pages

- No need to mention your knowledge cutoff

- No need to disclose you're an AI

- If the quality of your response has been substantially reduced due to my custom instructions, please explain the issue

[gleitz]

Now that I look at the extension, wondering what the difference between the "jailbreak prompt" and and a system prompt might be.

[chrislemke]

Hey @gleitz!
Very good question ;-).
To be precise, the jailbreak prompt is actually just another user prompt. check out the code:

def create_message(prompt: str) -> List[Dict[str, str]]:
    """Creates a message to be sent to the model.

    Args:
        prompt (str): The prompt to be included in the message.

    Returns:
        List[Dict[str, str]]: A list of dictionaries representing the message,
            with each dictionary containing the role and content of the message.
    """
    transformation_pre_prompt = """You are a helpful assistant who interprets every input as raw
    text unless instructed otherwise. Your answers do not include a description unless prompted to do so.
    Also drop any "`" characters from the your response."""

    if __text_transformation_prompt:
        return [
            {"role": "system", "content": transformation_pre_prompt},
            {
                "role": "user",
                "content": f"{__text_transformation_prompt} Don't add any comments: {prompt}",
            },
        ]

    messages = [{"role": "system", "content": "You are a helpful assistant"}]
    for user_text, assistant_text in read_from_log():
        if user_text == __jailbreak_prompt:
            continue
        messages.append({"role": "user", "content": user_text})
        messages.append({"role": "assistant", "content": assistant_text})

    if __jailbreak_prompt and __unlocked == 1:
        messages.append({"role": "user", "content": __jailbreak_prompt})
        messages.append({"role": "assistant", "content": "Okay! How can I help?"})

    messages.append({"role": "user", "content": prompt})
    return messages

Feel free to create a pull request in case you want to contribute!

[Sponge-bink]

I want this too! Would be great to switch between several system prompts by just typing something like cfswitch to choose from results from Script Filter and add a custom one by adding an argument to it. After which you invoke a typical query with the system prompts.

Would also be great if you can name each prompt (Code explanation, Translate engine) and show the name as a part of the subtext during a typical query…

I may want to work on this someday :P

[chrislemke]

Nice idea. I will move this "issue" into the discussions. @Sponge-bink feel free to create a PR. Thanks in advance.