Lightnews — Scholar-powered news

Steffen Herbold @sherbold.bsky.social · 11h

Very interesting paper that shows that LLMs are good. But not good enough. I wish the following from the conclusion would have made it into the abstract:

2 1

Steffen Herbold @sherbold.bsky.social · 6d

😂

Merriam-Webster @merriam-webster.com · 7d

We are thrilled to announce that our NEW Large Language Model will be released on 11.18.25.

Steffen Herbold @sherbold.bsky.social · 28d

Just accepted at TMLR:

We found evidence of copyright violations by LLMs even when we ask questions that were not part of the training. Indeed, we found that the amount of memorized content was independent from the questions being part of the training or not.

openreview.net/forum?id=ddo...

Studying memorization of large language models using answers to...

Large Language Models (LLMs) are capable of answering many software related questions and supporting developers by generating code snippets. These capabilities originate from training on massive...

openreview.net

2 6

Steffen Herbold @sherbold.bsky.social · 29d

This just in: Leading AI firm discovers confidence thresholds. More on this exciting development in news at 11.

openai.com/index/why-la...

(Honestly, OpenAI!?)

Why language models hallucinate

OpenAI’s new research explains why language models hallucinate. The findings show how improved evaluations can enhance AI reliability, honesty, and safety.

openai.com

1

Reposted by Steffen Herbold

Michael Dorner @michaeldorner.mastodon.social.ap.brid.gy · Sep 4

Scientific impact and achievement, redefined:
Huge congrats to #fraunhofer IIS on winning an #emmy for their JPEG XS compression standard 🏆🎉 […]

Original post on mastodon.social

mastodon.social

1 2

Steffen Herbold @sherbold.bsky.social · Sep 1

re

(I miss IRC)

(Now I feel old)

2

Steffen Herbold @sherbold.bsky.social · Aug 8

Dear all,

please enjoy your complementary "European Professor goes on Holiday" message.

See you in September.

Yours sincerely,
A European Professor

6

Reposted by Steffen Herbold

Hadas Kotek 🦄 @hadaskotek.bsky.social · Aug 8

Good news (for me!) my gender bias paper from 2023 still replicates with GPT-5.
Bad news (for everyone!) my gender bias paper from 2023 still replicates with GPT-5.
arxiv.org/pdf/2308.14921
hkotek.com/blog/gender-...

1 47 150

Steffen Herbold @sherbold.bsky.social · Aug 6

I wonder what my PhD students will think, once they discover that "someone" glued the three laws to the wall in the hallway. 🙃

PHD Comics @phdcomics.com · Aug 2

Newton's Laws of Graduation, Part 1

1 2

Reposted by Steffen Herbold

PHD Comics @phdcomics.com · Aug 4

Newton's Laws of Graduation, Part 2 - The Second Law

1 10 46

Reposted by Steffen Herbold

PHD Comics @phdcomics.com · Aug 6

Newton's Laws of Graduation, Part 3 - The Third Law 😆

3 9 46

Steffen Herbold @sherbold.bsky.social · Jul 29

Success, a luxury problem, and its solution:
🎉 Our quiz is a huge success and incredibly popular on YouTube with now over 100,000 views.
😐 We cannot answer all the feedback and comments individually anymore.
😀 We write a follow up article to answer the most important questions.

Uni Passau Research Magazine @unipassauresearch.bsky.social · Jul 29

The article on the reactions on YouTube is now available in English: www.digital.uni-passau.de/en/beitraege...

Quiz show ‘5 against AI’ receives huge response on YouTube

The German quiz show, in which a team of professors competes against AI, is causing lively discussion on YouTube. Answers to some of the questions.

www.digital.uni-passau.de

3

Reposted by Steffen Herbold

Johann-Mattis List @lingulist.de · Jul 23

It is official, our two long papers at #ACL2025 have now been published. Common work with Arne Rubehn (Concept Embeddings), and Frederic Blum and @sherbold.bsky.social (Automated Language Affiliation).

aclanthology.org/2025.acl-lon...
aclanthology.org/2025.acl-lon...

Partial Colexifications Improve Concept Embeddings

Arne Rubehn, Johann-Mattis List. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

aclanthology.org

1 4

Steffen Herbold @sherbold.bsky.social · Jul 21

My debut as TV-Show moderator - now live on Youtube.

We had a lot of fun with how the five professors answered questions on topics ranging from 90's music, counting peas, size of Asian countries, etc.

The only drawback: it is only available in German.

P.S. The humans won.

Uni Passau Research Magazine @unipassauresearch.bsky.social · Jul 21

Wie schlägt sich KI gegen professorale Expertise? Die Quiz-Show unter Moderation von @sherbold.bsky.social ist nun in voller Länge online. Wer sich vorab selbst mit der KI messen möchte, kann dies per Online-Quiz tun: www.digital.uni-passau.de/beitraege/20...

#KeepCALLM #5gegenKI

Quizshow 5 gegen KI - Professorenteam tritt gegen KI an

YouTube video by Universität Passau

www.youtube.com

1 1 5

Reposted by Steffen Herbold

Uni Passau Research Magazine @unipassauresearch.bsky.social · Jul 21

Wie schlägt sich KI gegen professorale Expertise? Die Quiz-Show unter Moderation von @sherbold.bsky.social ist nun in voller Länge online. Wer sich vorab selbst mit der KI messen möchte, kann dies per Online-Quiz tun: www.digital.uni-passau.de/beitraege/20...

#KeepCALLM #5gegenKI

Quizshow 5 gegen KI - Professorenteam tritt gegen KI an

YouTube video by Universität Passau

www.youtube.com

2 2

Steffen Herbold @sherbold.bsky.social · Jul 18

That was so much fun. I look forward to the video 😃

Uni Passau Research Magazine @unipassauresearch.bsky.social · Jul 18

Wer kann besser Erbsen zählen - 🤖oder unsere Profs?
Gestern bei uns im Studio: die Quizshow #5gegenKI

Moderator @sherbold.bsky.social hat mit elf Fragen immer wieder für Überraschung gesorgt. Eine Erkenntnis: Auch KI kann faul sein.

Die Show stellen wir in Kürze online. #KeepCALLM #staytuned

Wie viele Erbsen sind auf diesem Bild? Moderator Steffen Herbold, Professor für AI Engineering (links im Bild zu sehen) hat mit seinen Fragen nicht nur die Professoren und die Professorin auf der Bühne ins Grübeln gebracht. Auf der Bühne rätselten: Florian Lemmerich, Martina Padmanabhan, Johann-Mattis List, Moritz Müller und Brian Valerius.

4

Reposted by Steffen Herbold

Philipp Leitner @philippleitner.net · Jul 16

I'll just leave that quote here ...

2 1 2

Steffen Herbold @sherbold.bsky.social · Jul 16

Yesterday: Let's try to ground AI models in reality.
Now: Let's try to ground reality on AI models.

Fixes a lot of issues. I am impressed. 😅

They should call it LLM as a Physicists, then it gets accepted by the community ... right? (Looking at you, everybody trusting LLM as a judge!)

Steffen Herbold @sherbold.bsky.social · Jul 14

Happy to share that we published MAMUT @tmlrorg.bsky.social. We defined multiple data augmentation approaches to get more diverse mathematical data and show this improves pre-training.

Congrats to my student Jonathan Drechsel for his first publication! 🎉

www.fim.uni-passau.de/en/ai-engine...

Math Mutator (MAMUT) Accepted at TMLR

www.fim.uni-passau.de

1

Steffen Herbold @sherbold.bsky.social · Jul 11

No need, I can already feel the @icseconf.bsky.social paper bidding approaching 🙃

1

Steffen Herbold @sherbold.bsky.social · Jul 11

Starting the weekend on a Friday at 4pm with an empty inbox feels kind of strange. Good, but strange.

1 3

Reposted by Steffen Herbold

Roberto Verdecchia @robertoverdecchia.bsky.social · Jul 8

How much energy is needed to generate an image? 🎨🧠⚡️
Up to 4.08 Wh — like charging your phone to 40%!
In our new study we tested 17 models & 9,000+ runs.

Other key finds:
⚡️ Model energy use varies up to 46x
📐 Resolution matters, prompts don't
🛠️ Quantization ≠ savings

📄 Preprint: lnkd.in/dKWWAETW

2 3

Reposted by Steffen Herbold

Uni Passau Research Magazine @unipassauresearch.bsky.social · Jul 7

Können #LLMs einen neuen Zugang zum Recht eröffnen? Darüber spricht Brian Valerius, Professor für #KI im #Strafrecht, mit Rechtsanwalt Sven Galla, der KI bereits in der Praxis einsetzt.

📅 Donnerstag, 10. Juli, 18 Uhr
📍 Hörsaal 13

Mehr Infos: www.digital.uni-passau.de/generative-s...

#KeepCALLM

Ringvorlesung - Sprachmodelle: ein neuer Zugang zum Recht?

YouTube video by Universität Passau

youtu.be

1 1

Steffen Herbold @sherbold.bsky.social · Jul 4

The truly impressive thing about Zoom is that whenever they update the UI, it gets worse.

3

Steffen Herbold @sherbold.bsky.social · Jun 26

New pre-print: If you are wondering which models are good for non-code software engineering tasks, take a look at this work from my student Fabian Pena.

Also: Look at it if you want to know how to use Bayesian stats for ranking models.

arxiv.org/abs/2506.10833

Evaluating Large Language Models on Non-Code Software Engineering Tasks

Large Language Models (LLMs) have demonstrated remarkable capabilities in code understanding and generation; however, their effectiveness on non-code Software Engineering (SE) tasks remains underexplo...

arxiv.org

1