Tessera
Folgen

Niklas Stoehr

@niklasstoehr.bsky.social

Gemini Post-Training ⚫️ Research Scientist at Google DeepMind ⚫️ PhD from ETH Zurich

7 Posts1100 Follower218 Folgt
Alexander Hoyle@alexanderhoyle.bsky.social · 2mo ago

Very delighted to announce the next step in my career! After my postdoc at ETH, I will begin a joint appointment at TU Wien and the Complexity Science Hub Vienna as an Assistant Professor in NLP. I'm so grateful to all who helped me along the way And yes, I’m hiring! Details on PhD positions below

Photo of me in front of Stephansdom looking like a big dorkPhoto of main TU Wien buildingStock photo of Vienna for flavor
@nlpandcss.bsky.social · 4mo ago

✨The NLP+CSS workshop is returning to ACL 2026!✨ Our commitment deadline for pre-reviewed ARR submissions has been extended to April 5! Website/CfP: sites.google.com/site/nlpandc... #NLProc #CompSocialSci #ComputationalSocialScience #ACL2026NLP @aclmeeting.bsky.social

NLP+CSS Workshops - Call for Papers - 2026

Language is deeply intertwined with nearly all human social processes. We do not expect teenagers to speak like senior citizens, and we recognize the mutual dependency between language and the ways pe...

sites.google.com

Alexander Hoyle@alexanderhoyle.bsky.social · last yr.

Evaluating topic models (and document clustering methods) is hard. In fact, since our paper critiquing standard evaluation practices four years ago, there hasn't been a good replacement metric That ends today (we hope)! Our new ACL paper introduces an LLM-based evaluation protocol 🧵

Screenshot of first page of paper. It is here: https://arxiv.org/pdf/2507.00828 Abstract: Topic model and document-clustering evaluations either use automated metrics that align poorly with human preferences or require expert labels that are intractable to scale. We design a scalable human evaluation protocol and a corresponding automated approximation that reflect practitioners' real-world usage of models. Annotators -- or an LLM-based proxy -- review text items assigned to a topic or cluster, infer a category for the group, then apply that category to other documents. Using this protocol, we collect extensive crowdworker annotations of outputs from a diverse set of topic models on two datasets. We then use these annotations to validate automated proxies, finding that the best LLM proxies are statistically indistinguishable from a human annotator and can therefore serve as a reasonable substitute in automated evaluations
Paul Röttger@paul-rottger.bsky.social · last yr.

Are LLMs biased when they write about political issues? We just released IssueBench – the largest, most realistic benchmark of its kind – to answer this question more robustly than ever before. Long 🧵with spicy results 👇

Bild