How do you get LLMs to respond in a certain style? Prompt engineering often helps, but is not always enough, for example with sensitive data. This is where fine-tuning comes in: Instead of changing the prompt, the model itself is customised – locally and specifically.
In the workshop, we will show possible applications, carry out fine-tuning with our own data sets and explain which resources are required for this.
A laptop and access to Google Colab are required to take part. Ideally, you should bring along a subject area where you were dissatisfied with previous LLM answers.
This workshop will take place as part of the Practitioner Track at the international scientific conference “Communities and Technologies 2025”. Participants need to register at https://edih-swf.eu/event/finetuning-lokaler-llms/.
Language: German
- -
Practical information
- Where
- University of Siegen, Unteres Schloß 3, 57072 Siegen
- When
- -
- Who should attend
- SMEs, administrations and anyone interested in practical digitalization
- Organiser
- EDIH Südwestfalen
- Website
- https://edih-swf.eu/event/finetuning-lokaler-llms/