Intro to Prompt Engineering

Large Language Models (LLMs) have significantly impacted various fields, including higher education, advancing rapidly under the lead of commercial tech companies. However, many of these models are sensitive to the quality of user prompts, requiring well-crafted inputs to perform effectively. Studies have identified critical skills for LLM use, including understanding prompt structure, prompt literacy, and advanced methods such as few-shot prompting (providing examples of solved tasks), chain-of-thought prompting (encouraging step-by-step reasoning), and role modeling (assigning specific personas to tasks). This workshop will explore these techniques, highlight common pitfalls, and provide practical guidance to optimize the use of LLMs.

Literatur: Federiakin, D., Molerov, D., Zlatkin-Troitschanskaia, O., & Maur, A. (2024). Prompt engineering as a new 21st century skill. Frontiers in Education, 9, 1366434. https://doi.org/10.3389/feduc.2024.1366434 Sahoo, P., Singh, A. K., Saha, S., Jain, V., Mondal, S., & Chadha, A. A systematic survey of prompt engineering in large language models: Techniques and applications. Advance online publication. https://doi.org/10.48550/arXiv.2402.07927 Vatsal, S., & Dubey, H. A survey of prompt engineering methods in large language models for different NLP tasks. Advance online publication. https://doi.org/10.48550/arXiv.2407.12994

Methode: The learning outcomes of this session include gaining a formal theoretical perspective on prompting as a skill, becoming familiar with several effective prompting methods, and learning practical techniques to enhance model performance. The workshop methodology combines theoretical elements with interactive practical exercises. Participants will be encouraged to use an LLM model of their choice to observe how different prompting strategies influence the model’s behavior. Moreover, they will discuss how these prompt techniques work in different LLM models. The exercises will allow participants to experience firsthand how various prompting techniques can be more suitable for specific types of tasks, fostering a deeper understanding of their practical applications.