EUNIS Information Security Special Interest Group meets up regularly on a monthly basis to brainstorm on the current challenges of the information security field.
At the upcoming coffee break on 2 October 2026, Asbjørn Thorsen, InfoSec co-leader from Sikt, norway will explore how large language models (LLMs) can be benchmarked and evaluated from a security perspective. Asbjørn will share his approaches to testing LLM performance and vulnerabilities, designing meaningful evaluations, and interpreting and communicating the results. The session will also consider the challenges and limitations of current LLM security benchmarks.
When: 2 October 2026, 11.30 – 12.30 CEST, online
All EUNIS members are welcome to join. If you think this topic could be relevant or interesting to a colleague at your institution, feel free to invite them to attend as well.
