Applying prompts to GenAI/LLMs

Just a quick post on using ChatGPT4o to produce cartoons. The idea is inspired by a tweet by Dr Thomas Lancaster see below First, use a simple prompt to prime the system and generate the main panels. ChatGPT can now generate multiple images in a row. I asked for 8 panels. (2/6) pic.twitter.com/wqz2DyAUFe — Thomas Lancaster (@DrLancaster) June 17, 2024 Slightly modified it to get it to create the images and combine them in a GIF. Prompt 1 The language is British English. You are an experienced comic book designer and a witty writer. Create a guide to being a Computing student in Higher Education in the UK using a comic book narrative. This should be educational and entertaining. The comic will have 8 panels and the style consistent between panels, You will generate the graphics for each panel separately; as seperate images. using speech bubbles (a maximum of 5 words). Reduce all other textual output to a minimum.  Prompt 2 generate the images and combined the images into a file ca...

In a previous post  Ideas to project proposal 1  I discussed a relatively simple way to get GenAI to create a few ideas for a specific project and then create the start of a project proposal.  Now we are going to extend the idea and use Chain of Density Prompt to take the proposal and make it even richer. Chain of Density approach was discussed also in an earlier post  Improving your summarising A proposal is feed in and then an adapted forrm of the Chain of Density is used to refine the detail. The primary idea is that a richer in terms of information is produced. Prompt using the loaded file generate increasingly concise entity-dense summarises of the project proposal (now call it Article) . Sections are TITLE; INTRODUCTION ; STATEMENT OF THE PROBLEM ; PURPOSE OF THE STUDY ; ASSUMPTIONS AND HYPOTHESIS; DATA COLLECTION PROCEDURES; DATA ANALYSIS; DATA VALIDATION; ETHICAL ISSUES; REFERENCE LIST.   For each section:   Repeat 5 times, the following s...

One of the ways I use Generative AI with students is to take basic ideas for projects, usually a title, and get these tools to greater ideas and start of a project proposal. This is with all the usual caveats  Check the references (if any); It is going to be basic, so extend it. In this example I am going to use Co-pilot but the ChatGPT, etc can be used, employing a few basic prompt engineering basics: personas (who is the target audience?) and Templates (how do I want it to look?) to start this process. Example:  Project ideas for MSc Data Intelligence students (persona)  on a particular topic. The reply will include subheadings and relevant (hopefully) content for  TITLE, INTRODUCTION, PROBLEM STATEMENT. The prompt: " Taking the topic "Leveraging open-source tools to measure and present academics publications automatically from public domain data.". Give five innovative projects for a Master's level student dissertation in Data Intelligence. Each project example wi...

From AI to posts: Personally I think this one of the coolest quick wins with Generative AI getting it to produce social media posts (but not post them)  from either websites (the first video) or read in a conference paper and produce posts about it (second video). Gemini is great if you want to read in webpages. Transcript to potential blog post Sometimes you might produce a video (or use someone else's video) and turn the transcript into a blog post summarising it. For example a prompt such as  "Here is a transcript from a video please turn it in the text for a blog post of around 200 words suitable for a general audience." Consolidating websites A useful feature of Gemini is to read in multiple websites and consolidate their summarises into one. This might be a useful output in its own right, or a pre processing stage to do another activity (e.g get multiple insights on the summary. In essence the videos aim to show we can use Gen AI to do jobs for us. Please feel free...

Well sort of! There are approaches where we can get the generative AI to look at a problem from multiple perspectives (or personas) and bring the ideas generated, ideally informed by the others. to a final plan. One of the main strategy is called Tree of Thoughts (see here for more detail  https://www.forbes.com/sites/lanceeliot/2023/09/08/prompt-engineering-embraces-tree-of-thoughts-as-latest-new-technique-to-solve-generative-ai-toughest-problems/?sh=5ce79bdb2c8b ). The central idea is get a number of expert opinions, allow potential cross-fertilization of ideas, come up with actions or plans. Let see this action.  Scenario: Find out about the UK Government's plans on Disability support and then use Tree of Thoughts to produce some ideas for a company making disability equipment based on their website. Google's Gemini will be used. Stage 1 "UK Governments plans on Disability support ": Prompt:  Read, convert to plain text and consolidate information from the followi...

One of the most popular uses of Generative AI is summarising documents and information sources. Now for the but, could it be even better? Adams et (2023) the issues of using ChatGPT4 to produce the right amount of information in a summary. Balancing the detail in the summary without it becoming too dense, it is hard to follow, whilst ensuring it is relevant to the source. In the paper they investigated applying a technique they called "Chain of Density" (COD) to produce a better summary compared to a standard "summarise is this" prompt.   Chain of Density in Summary Start with the standard summarise this kind of prompt - fairly non-specific. 1. Identify one to three more informative entities from the article 2. Write a new denser summary of the same length including all the entities identified (don't remove entities when identified) expanding the detail Repeat steps 1 and 2 around 5 times. Applying guidelines around improving on the previous summary but must be ...

Two recent papers Eliot, L. (2023) and White, J.  et al.  (2023) have listed strategies for prompt engineering.  Taking Eliot, L (2023) paper categories different types of prompts to get different actions or outputs. The key point I feel is that LLMs are powerful tools that are a lot more than search tools, ie. "Tell me the answer to this". Selecting our prompts we can get LLMs to be closer to a tool for enhancing our creativity and productivity. Here's a table summarizing the prompt engineering strategies from Eliot (2023) and their summaries: Strategy Summary Imperfect Prompting Using intentionally imperfect prompts to generate creative or unexpected outcomes. Persistent Context and Custom Instructions Prompting Setting a persistent context or providing custom instructions to tailor AI responses. Multi-Persona Prompting Directing AI to adopt one or multiple personas for role-play or perspective exploration. Chain-of-Thought (CoT) Prompting Requesting AI to process ...