KAAI - Knowledge-Aware AI group @ScaDS.AI & TU Dresden
banner
kaaiscadsai.bsky.social
KAAI - Knowledge-Aware AI group @ScaDS.AI & TU Dresden
@kaaiscadsai.bsky.social
--Research group on LLMs and Knowledge Bases led by Prof. Simon Razniewski
Did you know that an LLM’s knowledge cutoff can be traced using dates in GPTKB?

For GPT-4.1, the declared cutoff in June 2024 closely matches the dates observed in GPTKB, down to the month level.

Deatils:https://arxiv.org/pdf/2510.07024

#DidYouKnowGPTKB
January 15, 2026 at 11:26 AM
We’re excited to share that a paper by Luca Giordano and Simon Razniewski, “Foundations of LLM Knowledge Materialization: Termination, Reproducibility, Robustness”, has been accepted to EACL 2026 Findings!

Read more about this paper here: arxiv.org/abs/2510.06780

#EACL2026 #EACL2026_Findings
Foundations of LLM Knowledge Materialization: Termination, Reproducibility, Robustness
Large Language Models (LLMs) encode substantial factual knowledge, yet measuring and systematizing this knowledge remains challenging. Converting it into structured format, for example through recursi...
arxiv.org
January 8, 2026 at 2:10 PM
As we welcome the New Year, Prof. Simon Razniewski from KAAI at @scadsai.bsky.social shares his wishes in a short talk.

🎉 We wish everyone a successful New Year, filled with innovation and curiosity.

Watch the video and join us in starting the year with a forward-looking perspective!
January 1, 2026 at 11:58 AM
KAAI group at @scadsai.bsky.social brings together researchers with complementary expertise, united by a shared commitment to advancing knowledge-aware and trustworthy artificial intelligence through collaborative research.
December 19, 2025 at 11:58 AM
This Tuesday, Prof.Simon Razniewski from KAAI group @scadsai.bsky.social & @tudresden.bsky.social gave a talk within the Faculty's Christmas lecture that explored LLMs through the eyes of Santa Claus, offering a refreshing perspective.
Wishing you all a thoughtful and joyful holiday season! 🎄✨🎊
December 17, 2025 at 8:23 AM
🤓Anyone can be anything
and apparently, anything can end up on gptkb.
Even Zootopia 2.
👇click on
gptkb.org/entity/Zooto...
to find more about gptkb, our exciting project in KAAI group at @scadsai.bsky.social & @tudresden.bsky.social 🥳
December 15, 2025 at 1:37 PM
Hello,
we are the Knowledge-aware AI research group at @scadsai.bsky.social & @tudresden.bsky.social.
We do research in language models and knowledge bases. We regularly publish at premier AI and NLP venues like ACL, EMNLP and AAAI.
💡Our signature project is GPTKB - check it out at
👉 gptkb.org !
December 12, 2025 at 8:55 AM