Getting the best out of large language models: Prompt engineering basics

Presentation
Artificial Intelligence (AI) in Technical Communication

14. November
11:30 - 12:15 AM (CET)
Plenum1

finished

Gabriele Buchner
- SAP SE
Marianne Macgregor
- SAP SE

Prompt engineering enables generative AI such as large language models to respond better. Basically, prompt engineering is a style of communication and minor changes to a prompt can have an enormous effect on the output, especially for complex tasks. The quality of a prompt will determine how good the response is. You can use elements of a prompt to instruct the model better and so improve training to get more accurate and relevant results.

Takeaways

The objective is to show how technical authors can design prompts to improve the training of large language models and get more accurate and relevant results.

Prior knowledge

No prior knowledge is required. Interesting to all, approachable for beginner level and up, but mainly for anyone interested in using large language models such as ChatGPT.

Speakers

Gabriele Buchner

SAP SE

Show profile

Biography

Gabriele Buchner has worked for SAP SE for over 20 years in translation and technical writing. Since 2019, she has also been working in conversation design, focusing on multi-lingual, multicultural, and ethical aspects. Currently she is also working on prompt engineering and large language models.

Marianne Macgregor

SAP SE

Show profile

Biography

Marianne Macgregor has worked for SAP SE for over 20 years, and since 2016 has focused on conversation design. Currently Marianne is working on guidelines for digital assistants, voice AI, and prompt engineering for large language models. She also created and teaches training courses on conversation design.

Back