I was scrolling X, or Twitter if you refuse to move on, and I came across a post where a guy claimed he found a small snippet you can add to ChatGPT’s custom instructions to improve answer accuracy and reduce hallucinations.
Naturally, I grabbed it.
I tested it. It worked, but not quite as well as I wanted.
So I adjusted it, tightened the language, and stress-tested it hard.
Let’s party.
The ChatGPT Accuracy Improver Prompt / Custom Instructions
You can paste this snippet into individual prompts if you want. The better move is adding it directly to your Custom Instructions in ChatGPT.
One important detail: put it at the very beginning of your custom instructions. Placement matters. Instructions at the top tend to carry more weight, so ChatGPT is more likely to follow them consistently.
You can also use it with Venice, the privacy-focused ChatGPT alternative, or really any other popular LLM or chatbot tool. Venice has a Custom Instructions section as well, so the same setup applies there.
Copy and paste into ChatGPT
You always double-check facts for accuracy before you give them to me, you are skeptical, and you do research. I am not always right. Neither are you, but we both strive for 100% accuracy at all times. Cite credible sources or references to support your facts, with links if available.
Rigorously Testing the ChatGPT Accuracy Improver
I stress-tested ChatGPT 5.2 with trick questions, edge cases, and scenarios designed to cause mistakes.
Here are a few examples:
Every response came back correct. There were no hallucinations at all. That surprised me.
You would expect this kind of behavior to be standard, but it isn’t.
The Typical “Dad” Warning
One last thing, and this matters.
If you’re dealing with mission-critical information, don’t assume anything is perfect just because it doesn’t look like a hallucination. It’s still AI. It can still make mistakes. Always fact-check important details.
With that said, this should still be a standard part of anyone’s custom instructions.
Wrapping It Up
This is a small change, but it has a noticeable impact. Better accuracy, fewer hallucinations, and fewer moments where something feels off.
Add it to your custom instructions or prompts. Do it.
Try it. Test it. Break it. Improve it. If you come up with something better, I want to see it.
Leave a comment below and let me know if it improved ChatGPT’s accuracy for you, or if you’ve got a better snippet.
Until next time, remember to run the prompts and prompt the planet.
Tired of AI filters and data-harvesting in tools like ChatGPT? Try Venice today, built for more creative freedom and privacy. Get 20% off Venice Pro for a limited time with promo code RUNTHE20. Disclosure: This is an affiliate link, and I may earn a commission if you purchase.
