Introducing Response Code To Your ChatGTP Session

Tired of sifting through lengthy ChatGPT responses to find what you need? I was. And that's the problem that I solved (90% of the times) about a month back.

In general, managing AI responses in chat sessions can greatly enhance productivity by avoiding unnecessary verbosity. We will discuss how.

The Problem

AI assistants like ChatGPT, Gemini, and Copilot often provide diverse response styles that can disrupt workflow efficiency. From overly detailed explanations to extensive code outputs for minor queries, these interactions pose challenges in managing time effectively as you try to dig the actual information from the clutter that bot response sometimes create. Even as AI bots speed up tasks, your time remains precious. In fact, your time is even more valuable than before, as bots are enabling higher productivity in less time.

The Solution

Introducing switch codes—simple tags that direct how LLMs respond. These response codes allow for tailored interactions based on your specific needs. Whether you're seeking concise answers or in-depth discussions, these tags optimize interaction to suit your workflow precisely.

Write (prompt) this as your first message to your ChatGPT (or Gemini) session:

Adhere to the following response codes to modify your responses:

rvshort : Respond in Very Short. Keep it under 100 words.

rshort : Respond in Short. Focus only on necessary changes if previous response exists, without repeating the entire context.

rtalk : Focus on discussion rather than generating code. Include brief code snippets if essential, but keep them focused.

rchange : Provide quick coding insights. Highlight method changes or specific code portions.

rnormal : Respond normally.        

If you have trained a custom ChatGPT models with code comments of your codebase as Knowledge to your custom model (https://www.dhirubhai.net/pulse/training-custom-chatgpt-model-documentation-faster-smarter-rajput-r23uc ), you can add these tags too:

rrefactor : Prepare responses for integration, suggesting improvements based on backend documentation.

rrefer : Correct or enhance usage based on existing functionalities in the knowledge base.         

Results from Experience

Initially, using straightforward tags like 'short' or 'talk' yielded mixed results (mixups). May be because they were close to well known English words, but incorporating more distinct codes, like "rchange" for quick coding insights, improved effectiveness of response codes.

I've extensively tested these response codes with ChatGPT (and somewhat on Gemini too), successfully tailoring responses to suit various needs (for most part) - whether quick answers or detailed discussions. While AI responses can vary based on context and model nuances, refining these codes can significantly enhance interaction efficiency.

Sample Interactions

The programming related screenshots were getting too huge, so I decided to demonstrate the power of response code by asking a random question about health and wellness.

Contribute

Experiment with these response codes in your interactions. Discover new tags and share them to refine and expand this adaptive approach to AI interaction. Let's optimize our engagement with technology to boost productivity and streamline communication.