Lightnews — Scholar-powered news

Reposted by Bastian Bunzeck

Maria Antoniak @mariaa.bsky.social · 2h

Keynote at #COLM2025: Nicholas Carlini from Anthropic

"Are language models worth it?"

Explains that the prior decade of his work on adversarial images, while it taught us a lot, isn't very applied; it's unlikely anyone is actually altering images of cats in scary ways.

1 15 47

Reposted by Bastian Bunzeck

thebes @vgel.me · 2d

i wrote a custom llm sampler for llama-3.1-8b so it could only say words that are in the bible

12 62 380

Reposted by Bastian Bunzeck

Marianne de Heer Kloots @mdhk.net · 6d

Huge congrats to the envisionBOX team for the Open Science award nomination! 🎉

My tutorial on speech analysis tools in Python from the Unboxing Multimodality summer school (github.com/mdhk/unboxin...) is now also available at envisionbox.org

Thanks for the invitation & this great initiative! 👏

wimpouw @wimpouw.bsky.social · 7d

www.envisionbox.org has been shortlisted for the Leo Waaijers Open Science price: ukb.nl/en/news/shor...

@babajideowoyele.bsky.social @jamestrujillo.bsky.social @sarkadava.bsky.social @DavideAhmar @acwiek.bsky.social

Amazing Markus Küpper made an animated video:
www.youtube.com/watch?v=HduI...

EnvisionBOX overview2025

YouTube video by Wim Pouw

www.youtube.com

3 5

Reposted by Bastian Bunzeck

Stefan Hartmann @stefanhartmann.bsky.social · 7d

Gentle reminder that the #CfP for #Evolang2026 @evolangconf.bsky.social is still open - deadline October 26! sites.google.com/york.ac.uk/e...

EVOLANG 2026 - Call for Papers

sites.google.com

3 4

Reposted by Bastian Bunzeck

Aaron Mueller @amuuueller.bsky.social · 8d

What's the right unit of analysis for understanding LLM internals? We explore in our mech interp survey (a major update from our 2024 ms).

We’ve added more recent work and more immediately actionable directions for future work. Now published in Computational Linguistics!

2 14 38

Reposted by Bastian Bunzeck

Gemma Boleda @gboleda.bsky.social · 9d

New paper! 🚨 I argue that LLMs represent a synthesis between distributed and symbolic approaches to language, because, when exposed to language, they develop highly symbolic representations and processing mechanisms in addition to distributed ones.
arxiv.org/abs/2502.11856

Sigmoid function. Non-linearities in neural network allow it to behave in distributed and near-symbolic fashions.

1 11 26

Reposted by Bastian Bunzeck

Quanta Magazine @quantamagazine.bsky.social · 9d

Many AI researchers draw inspiration from neuroscience. Naomi Saphra favors a different analogy. Interpretability, in her view, should take a cue from evolutionary biology.

To Understand AI, Watch How It Evolves | Quanta Magazine

Naomi Saphra thinks that most research into language models focuses too much on the finished product. She’s mining the history of their training for insights into why these systems work the way they…

www.quantamagazine.org

1 10 36

Bastian Bunzeck @bbunzeck.bsky.social · 13d

My very first book review is out now 📚
Muchas gracias to @stefanhartmann.bsky.social for inviting me, looking forward to our next project(s) 😇

Stefan Hartmann @stefanhartmann.bsky.social · 13d

It's been a while since I've written a book review - here's our review of Herbst & Hoffmann (2024), my first but definitely not last collaboration with the brilliant @bbunzeck.bsky.social doi.org/10.1017/S136... ($)

Thomas Herbst and Thomas Hoffmann, A Construction Grammar of the English language: CASA – a constructionist approach to syntactic analysis (Cognitive Linguistics in Practice 5). Amsterdam and Philadel...

doi.org

1 8

Reposted by Bastian Bunzeck

Pranav A @pranav-nlp.bsky.social · 14d

I'm conducting research on how ACL's peer review policies impact NLP research quality, career trajectories, and inclusivity within our community. I am running a survey, which would take around 7-10 mins to complete: forms.cloud.microsoft/e/j2jr9nH3X0

I would really appreciate insights from y'all!

I'm conducting research on how ACL's peer reviewing policies impact NLP research quality, career trajectories, and inclusivity within our community. Your insights—whether you're a seasoned reviewer, early-career researcher, or anywhere in between—are invaluable.
The survey takes 7-10 minutes and covers topics like review quality, reviewer assignment, and accessibility barriers. All responses are confidential and will help inform evidence-based improvements to our peer review processes.

1 6 6

Reposted by Bastian Bunzeck

Anne Lauscher @a-lauscher.bsky.social · 15d

🚨 Are you looking for a PhD in #NLProc dealing with #LLMs?
🎉 Good news: I am hiring! 🎉
The position is part of the “Contested Climate Futures" project. 🌱🌍 You will focus on developing next-generation AI methods🤖 to analyze climate-related concepts in content—including texts, images, and videos.

1 14 22

Reposted by Bastian Bunzeck

Daniel Duran @simphon.bsky.social · 16d

Attending the The Second International Workshop on Construction Grammars and NLP (CxGs+NLP 2025) in Düsseldorf, Germany? Check out the poster “Do Construction Distributions Shape Formal Language Learning In German BabyLMs?” by Bastian Bunzeck and colleagues! @bbunzeck.bsky.social #CRC1646 #LINCC

1 3

Bastian Bunzeck @bbunzeck.bsky.social · 17d

From conference to conference: September ends with a trip to #IWCS in beautiful Düsseldorf. Hyped for two days of semantics (and two more days of construction grammar and NLP). 🥳

1 10

Reposted by Bastian Bunzeck

Stefan Hartmann @stefanhartmann.bsky.social · 20d

The first of the three corpora of German-English bilingual children's early speech that we've been working on for the last few years is finally publicly available! 🥳 🎉 talkbank.org/childes/acce...

CHILDES English-German MPI-EVA-Leipzig Corpus

talkbank.org

1 14 47

Reposted by Bastian Bunzeck

Daniel Duran @simphon.bsky.social · 27d

“Developmentally plausible pretraining, now also auf Deutsch: a BabyLM Dataset for German” — Today I had the pleasure to present our German BabyLM dataset together with the first author Bastian Bunzeck @bbunzeck.bsky.social‬ to an interested and engaging audience at #KONVENS2025 in Hildesheim.

Daniel Duran and Bastian Bunzeck at the poster presentation

3 9

Bastian Bunzeck @bbunzeck.bsky.social · 28d

Our BabyLMs at #konvens 🥳

CLAUSE - Computational Linguistics @ Bielefeld University @clausebielefeld.bsky.social · 28d

Happening now: Sina‘s keynote on our BabyLM work. 🥳

4

Bastian Bunzeck @bbunzeck.bsky.social · 29d

From conference to conference — after last week’s #semdial I am at #konvens in Hildesheim this week. I will be presenting out German BabyLM Corpus (with @simphon.bsky.social) and our PI Sina Zarrieß will give a Keynote on BabyLMs tomorrow. 🥳

1 6

Reposted by Bastian Bunzeck

Max Planck Institute for Psycholinguistics @mpi-nl.bsky.social · Sep 9

Children are not the main agents of language change. New paper by @limorraviv.bsky.social @damianblasi.bsky.social & @vkempe.bsky.social.
psycnet.apa.org/doi/10.1037/...
www.mpi.nl/publications...

10 28

Reposted by Bastian Bunzeck

Adele Goldberg @adelegoldberg.bsky.social · Sep 7

Abstract deadline changed to *December 1, 2025*

Adele Goldberg @adelegoldberg.bsky.social · Aug 24

UPDATE: Abstract deadline: Nov 1, 25! Invited speakers: Corrine Occhino, Dagmar Divjak, Idan Blank, Randy Allen Harris, Gary Lupyan, Laura Michaelis, Kanishka Misra !
@dagmardivjak.bsky.social @randyallenharris.bsky.social @congramqueen.bsky.social @glupyan.bsky.social @kanishka.bsky.social

Adele Goldberg @adelegoldberg.bsky.social · Aug 15

📌 👉 The 14th International Construction Grammar conference will be held at Princeton, June 4-7, 2026

Usage-based analyses and Empirical methods

Stay tuned for updates!

5 13

Reposted by Bastian Bunzeck

Daniel Duran @simphon.bsky.social · Sep 5

It's a wrap! Thanks to the organizers, presenters and all participants for an inspiring and engaging #semdial2025 #bialogue conference at Bielefeld University! I had fun.
In case you missed it, proceedings can be found on the website:
semdial2025.github.io

Members of the Local Organizing Committee at the closing of SemDial 2025.

2 3

Bastian Bunzeck @bbunzeck.bsky.social · Sep 5

🗣️🗣️🗣️❗️❗️❗️

1 2

Reposted by Bastian Bunzeck

Daniel Duran @simphon.bsky.social · Sep 3

“Child-directed speech is fine-tuned to children’s developmental needs” — @bbunzeck.bsky.social from the A02 project of #CRC1646 #LINCC, presented a great poster today at #semdial2025 #bialogue based on earlier work in collaboration with Holger Diessel.

Bastian Bunzeck in front of his poster "Child-directed speech is fine-tuned to children’s developmental needs (Bastian Bunzeck, Holger Diessel)"

1 4

Bastian Bunzeck @bbunzeck.bsky.social · Sep 3

Check out the paper here: journals.sagepub.com/doi/10.1177/...

2

Bastian Bunzeck @bbunzeck.bsky.social · Sep 3

I will present a poster on the First Language article I wrote with Holger Diessel now at #semdial 😁💬

1 1 3

Reposted by Bastian Bunzeck

CLAUSE - Computational Linguistics @ Bielefeld University @clausebielefeld.bsky.social · Sep 3

Leonie Schade asks whether it takes two to do an articulatory tango 😁

1 1 6

Bastian Bunzeck @bbunzeck.bsky.social · Sep 3

Now coming up: session 1 on naturalistic dialogue 👌

3