With the recent advancements in the field of Natural Language Processing (NLP), Generative AI (Gen-AI) tools are rapidly evolving and have already become powerful tools for various sectors. Including the widely known tool ChatGPT, there are various competitors in the market with recently proposed open-source alternatives. Depending on the method under the hood and the data set genAI are trained on, varying tools can have certain capabilities and limitations. To have a better conservation with these Gen-AI tools and fully harness their potential, it is crucial to be aware of the concept of prompt engineering and power.
This workshop aims to introduce the main strategies of prompt engineering including tips and certain practices to achieve optimal outputs with Gen-AI tools. After providing a hands-on overview of prompt engineering, certain features of ChatGPT will be discussed for fostering the research practices. The target audience for this workshop is mainly researchers and practitioners working in different fields who are interested in using the Gen-AI tools more efficiently and responsibly in their daily work.
This is a beginner-friendly workshop. No previous knowledge on the topic is required/expected and the trainer will cover the basics of the method. If you want to explore the topic before the workshop you can have a look at this Prompt Engineering Guide.
- Understand the fundamentals of Gen-AI and prompt engineering; explain how prompts influence the performance of Gen-AI models.
- Identify the strengths and limitations of Gen-AI tools; recognise how underlying training data and model design affect outputs in research contexts.
- Apply core prompting strategies to generate more accurate and useful results.
By attending this course, you will become familiar with the following skills:
- Different prompt engineering strategies (e.g., zero-shot, few-shot, chain-of-thought) and how to apply them effectively.
- The core mechanism behind modern LLMs, including attention, embeddings, and encoder/decoder structures.
