## Prompt resource links - [[26 prompt hacks for ChatGPT according to science]] - [Job Description Prompt](https://docs.google.com/document/d/1u00QiirBtOtZhXJgay10oH4gjWQ7-iEdWWnt5YstBFw/edit) from Dan Shapiro, via [[Ethan Mollick]] ([link](https://www.oneusefulthing.org/p/latent-expertise-everyone-is-in-r?publication_id=1180644&utm_campaign=email-post-title&r=3nfnew&utm_medium=email)) - [Edward Frank Morris on LinkedIn: The Renaissance of Prompting: Part 1 | 175 comments](https://www.linkedin.com/posts/edwardfmorris_the-renaissance-of-prompting-part-1-ugcPost-7214938473005264898-VNHd?utm_source=combined_share_message&utm_medium=member_ios) - [MacStories' ChatGPT Proofreading Prompt](https://club.macstories.net/posts/macstories-chatgpt-proofreading-prompt) ## Open AI’s prompt documents - [How do I create a good prompt for an AI model like GPT-4? \| OpenAI Help Center](https://help.openai.com/en/articles/4936848-how-do-i-create-a-good-prompt-for-an-ai-model-like-gpt-4) - [OpenAI Platform - Reasoning Best Practices](https://platform.openai.com/docs/guides/reasoning-best-practices) - [OpenAI Platform - Prompt Engineering](https://platform.openai.com/docs/guides/text?api-mode=chat#prompt-engineering) ## BRIDGE method My colleague Jared Reimer taught me this formula for creating a prompt:   B.R.I.D.G.E.   BACKGROUND: What the model needs to know (context).  This can include your role, industry, audience, objective, constraints, or any relevant background information.   REQUEST: The core action or what you want the model to do.  Clear, specific requests generally lead to clearer, more specific outputs.   INPUT: Your data or sources. This is where you provide the information you want the model to use and optionally specify what it should not use.  It’s important to follow SRS best practices around data privacy at this step.   DELIVERABLES: What the output should look like.  Define the format, level of detail, structure, or length of what you’re asking the model to produce.   GUARDRAILS: What to avoid.  Specify what the model should not do like making assumptions, introducing unrelated content, or citing sources you didn’t provide.  It was recommended to maintain a standard set of guardrails for common request types.   EVALUATION: How to check quality.  Define how the model should review its own work like calling out assumptions, data gaps, added logic, limitations, or risks in the output. ## Sample prompts ![[AI Tip - Provide a list of installed apps]] --- Source unknown: > "Ask me one question at a time until you have enough information to complete the task." --- From [[Joe Elmendorf]] via LinkedIn" >I am going to ask you a question. You are a reference librarian so instead of immediately answering the question you are going to ask probing questions to get at what it is that I am really interested in knowing - one at a time, using the answer from each as an input to the next - consistent with what a reference librarian would ask, to gain context about my question. Repeat this pattern of asking a question, receiving an answer, and asking a followup question based on the answer at least five times before assembling your authoritative answer to the actual question I am asking. --- From [[Co-Intelligence - Living and Working with AI]], by Ethan Mollick Prompt: >Think this through step by step: come up with good analogies for an AI tutor. First, list possible analogies. Second, critique the list and add three more analogies. Next, create a table listing pluses and minuses of each. Next, pick the best and explain it. Advice >A lot of active research is happening around the best way to “program” an LLM, but one practical implication is that it can help to give the AI explicit instructions that go step by step through what you want. One approach, called chain-of-thought prompting, gives the AI an example of how you want it to reason, before you make your request. Even more usefully, you can also provide step-by-step instructions that build on each other, making it easier to check the output of each step (letting you refine the prompt later), and which will tend to make the output of your prompts more accurate. > --- From [Ethan Mollick on LinkedIn](https://www.linkedin.com/posts/emollick_a-really-useful-prompt-for-writing-review-activity-7367024966216724480-71pV?utm_source=share&utm_medium=member_ios&rcm=ACoAAAArP08BMtkDTuy0V840e_XuwGnahY9Swkc) >A really useful prompt for writing: "review this for accuracy, look up any facts you may want to challenge or explore." <iframe style="border: 0; width: 100%; height: 450px;" allowfullscreen frameborder="0" src="https://raindrop.io/stumaxis/ai-prompt-engineering-45111301/embed"></iframe> --- **Relates to**: [[AI]], [[ChatGPT]], [[GenAI prompt resources]]