ChatGPT and generative AI models like it are all the rage, but users often write prompts for these models that are less effective than they could be. Based on the extensive documentation provided by OpenAI and interviews with key OpenAI staff on how large language models are constructed, these are the principal components prompts should have for maximum effectiveness: Role, Action, Context, Execute, or RACE. In general, longer, more detailed prompts are better, up to a limit of about 3,000 words. To maximize the impact of your prompts, use the following cheat sheet to write the best, most effective prompts possible. Click on the image or the link to get the PDF version.
Want to learn more about how systems like ChatGPT work? Watch this talk!
Want some help implementing AI within your organization? Contact us!