You now have a pretty solid understanding of Large Language Models and their underlying mechanisms.
You most likely use LLMs in your daily work.
But what are LLMs good for and what should you not do with them?
Explore the limitations of LLMs.
See the subsequent slides for example scenarios
Scenario: Seminar Paper.
Let the LLM generate a fictional seminar paper.
How is the quality of the reference section?
What about the citation style?
Is all metadata complete?
Do all references exist?
Scenario: Program Code.
(Re-)visit the tasks from the STAP seminar.
Are the solutions correct?
How specific are the solutions?
Do the solutions account for a given framework?