The Ultimate ChatGPT Prompt Hack for Data Scientists

In the fast-paced world of machine learning and artificial intelligence, discovering innovative methods to leverage available tools can significantly streamline our tasks and produce improved results. Recently, I chanced upon an episode of the Super Data Science Podcast, hosted by Jon Krohn. In this episode, Jon discusses an invaluable tip from Shaan Khosla, a recognized expert in large language model prompts, on how to get enhanced responses from ChatGPT.

The Genesis of the Prompt

Jon Krohn, host of the Super Data Science Podcast, shared a particularly insightful tip he learned from Shaan Khosla, a prompt engineer at Nebula. Shaan, having hands-on experience with both proprietary and open-source large language models, has mastered the art of ChatGPT’s Custom Instructions. With these instructions, one can automatically fetch more improved responses from the model.

How to Implement

In case you haven’t explored this feature yet, ChatGPT offers Custom Instructions to customize the bot’s responses. To access this:

1. Navigate to the bottom left-hand corner in ChatGPT, where your email address is located.
2. Click on the three dots and select ‘Custom Instructions’.
3. Two fields appear. The first asks for information about the user for better responses, while the second pertains to how you’d like ChatGPT to respond.

While there’s an option to fill in details about yourself, the hack makes use of the second field: “how would you like ChatGPT to respond?”

The Magic Prompt

Here’s the custom prompt that Shaan designed:

“I need you to help me with a task. To help me with a task, first, come up with a detailed outline of how you think you should respond. Then critique the ideas in this outline, mention the advantages, disadvantages, and ways it could be improved. Then use the original outline and the critiques you made to come up with your best possible solution. Overall, your tone should not be overly dramatic. It should be clear, professional, and direct. Don’t sound robotic or like you’re trying to sell something. You don’t need to remind me, you’re a large language model. Get straight to what you need to say to be as helpful as possible.”

What sets this prompt apart? Shaan emphasizes breaking down tasks and critiquing initial responses. By prompting the model to critique its own initial ideas, it iterates to produce, in many cases, far superior outputs than its original inclinations.

Moreover, for professionals such as data scientists, AI researchers, and software engineers, there’s a desire for clear, concise, and professional responses. Shaan’s prompt ensures this by directing ChatGPT to bypass its often ‘bubbly’ nature and get straight to the point.

Concluding Thoughts

With the ever-growing influence of large language models in our professional tasks, such tricks and tips can be game-changers. If you’re keen on diving deeper or visualizing this hands-on, Jon Krohn has provided a YouTube version of this episode. For those looking to further optimize their interactions with ChatGPT, Shaan Khosla’s prompt is a worthy addition to your toolkit.