Lightnews — Scholar-powered news

Reposted by Leonie Weissweiler

Kanishka Misra 🌊 @kanishka.bsky.social · 2d

Happening now! Poster 42!

1 4

Reposted by Leonie Weissweiler

Kanishka Misra 🌊 @kanishka.bsky.social · 4d

Traveling to my first @colmweb.org🍁

Not presenting anything but here are two posters you should visit:

1. @qyao.bsky.social on Controlled rearing for direct and indirect evidence for datives (w/ me, @weissweiler.bsky.social and @kmahowald.bsky.social), W morning

Paper: arxiv.org/abs/2503.20850

Both Direct and Indirect Evidence Contribute to Dative Alternation Preferences in Language Models

Language models (LMs) tend to show human-like preferences on a number of syntactic phenomena, but the extent to which these are attributable to direct exposure to the phenomena or more general propert...

arxiv.org

1 5 13

Leonie Weissweiler @weissweiler.bsky.social · 25d

I'll be dreaming of them once Swedish winter starts...

Leonie Weissweiler @weissweiler.bsky.social · 25d

📢Life update📢

🥳I'm excited to share that I've started as a postdoc at Uppsala University NLP @uppsalanlp.bsky.social, working with Joakim Nivre on topics related to constructions and multilinguality!

🙏Many thanks to the Walter Benjamin Programme of the DFG for making this possible.

3 2 28

Reposted by Leonie Weissweiler

Jaap Jumelet @jumelet.bsky.social · Aug 1

Happening now at the SIGTYP poster session! Come talk to Leonie and me about MultiBLiMP!

1 2 20

Leonie Weissweiler @weissweiler.bsky.social · Jul 23

Congratulations! 🥳 (both to you and to all us Germans 😊)

1 1

Leonie Weissweiler @weissweiler.bsky.social · Jun 21

Hi #NLP community, I'm urgently looking for an emergency reviewer for the ARR Linguistic Theories track. The paper investigates and measures orthography across many languages. Please shoot me a quick email if you can review!

5 5

Leonie Weissweiler @weissweiler.bsky.social · Jun 11

I'm looking for a reviewer for a paper on measuring syntactic productivity (lots of maths!) due a week from now. Please shoot me an email if you could review!

3

Leonie Weissweiler @weissweiler.bsky.social · Jun 9

@jlibovicky.bsky.social

Leonie Weissweiler @weissweiler.bsky.social · Jun 3

🥳Congratulations! I'm so excited for you but also sad to miss you (again)!

1 1

Reposted by Leonie Weissweiler

Ai2 @ai2.bsky.social · May 9

Do LLMs learn language via rules or analogies?
This could be a surprise to many – models rely heavily on stored examples and draw analogies when dealing with unfamiliar words, much as humans do. Check out this new study led by @valentinhofmann.bsky.social to learn how they made the discovery 💡

Valentin Hofmann @valentinhofmann.bsky.social · May 9

Thrilled to share that this is out in @pnas.org today! 🎉

We show that linguistic generalization in language models can be due to underlying analogical mechanisms.

Shoutout to my amazing co-authors @weissweiler.bsky.social, @davidrmortensen.bsky.social, Hinrich Schütze, and Janet Pierrehumbert!

Valentin Hofmann @valentinhofmann.bsky.social · Dec 5

📢 New paper 📢

What generalization mechanisms shape the language skills of LLMs?

Prior work has claimed that LLMs learn language via rules.

We revisit the question and find that superficially rule-like behavior of LLMs can be traced to underlying analogical processes.

🧵

1 5 22

Reposted by Leonie Weissweiler

Valentin Hofmann @valentinhofmann.bsky.social · May 9

Thrilled to share that this is out in @pnas.org today! 🎉

We show that linguistic generalization in language models can be due to underlying analogical mechanisms.

Shoutout to my amazing co-authors @weissweiler.bsky.social, @davidrmortensen.bsky.social, Hinrich Schütze, and Janet Pierrehumbert!

Valentin Hofmann @valentinhofmann.bsky.social · Dec 5

📢 New paper 📢

What generalization mechanisms shape the language skills of LLMs?

Prior work has claimed that LLMs learn language via rules.

We revisit the question and find that superficially rule-like behavior of LLMs can be traced to underlying analogical processes.

🧵

1 6 37

Reposted by Leonie Weissweiler

Kwanghee Choi @juice500ml.bsky.social · Apr 29

Can self-supervised models 🤖 understand allophony 🗣? Excited to share my new #NAACL2025 paper: Leveraging Allophony in Self-Supervised Speech Models for Atypical Pronunciation Assessment arxiv.org/abs/2502.07029 (1/n)

2 10 15

Reposted by Leonie Weissweiler

Verena Blaschke @verenablaschke.bsky.social · Apr 29

On my way to #NAACL2025 where I'll give a keynote at the noisy text workshop (WNUT), presenting some of the challenges & methods for dialect NLP + also discussing dialect speakers' perspectives!

🗨️ Beyond “noisy” text: How (and why) to process dialect data
🗓️ Saturday, May 3, 9:30–10:30

1 7 27

Leonie Weissweiler @weissweiler.bsky.social · Apr 25

self-promotion, but we argued similar things here: bsky.app/profile/weis...

Leonie Weissweiler @weissweiler.bsky.social · Feb 20

✨New paper✨

Linguistic evaluations of LLMs often implicitly assume that language is generated by symbolic rules.
In a new position paper, @adelegoldberg.bsky.social, @kmahowald.bsky.social and I argue that languages are not Lego sets, and evaluations should reflect this!

arxiv.org/pdf/2502.13195

1 3

Leonie Weissweiler @weissweiler.bsky.social · Apr 7

🌍📣🥳
I could not be more excited for this to be out!

With a fully automated pipeline based on Universal Dependencies, 43 non-Indoeuropean languages, and the best LLMs only scoring 90.2%, I hope this will be a challenging and interesting benchmark for multilingual NLP.

Go test your language models!

Jaap Jumelet @jumelet.bsky.social · Apr 7

✨New paper ✨

Introducing 🌍MultiBLiMP 1.0: A Massively Multilingual Benchmark of Minimal Pairs for Subject-Verb Agreement, covering 101 languages!

We present over 125,000 minimal pairs and evaluate 17 LLMs, finding that support is still lacking for many languages.

🧵⬇️

1 13

Leonie Weissweiler @weissweiler.bsky.social · Mar 31

We can use small LMs to test hypotheses about the language network and how everything is connected!

Here, we find that dative alternation preferences are learned from dative-specific input statistics *and* from more general short-first preferences.

Great work by @qyao.bsky.social, go follow him!

Qing Yao @qyao.bsky.social · Mar 31

LMs learn argument-based preferences for dative constructions (preferring recipient first when it’s shorter), consistent with humans. Is this from memorizing preferences in training? New paper w/ @kanishka.bsky.social , @weissweiler.bsky.social , @kmahowald.bsky.social

arxiv.org/abs/2503.20850

examples from direct and prepositional object datives with short-first and long-first word orders:
DO (long first): She gave the boy who signed up for class and was excited it.
PO (short first): She gave it to the boy who signed up for class and was excited.
DO (short first): She gave him the book that everyone was excited to read.
PO (long-first): She gave the book that everyone was excited to read to him.

7

Leonie Weissweiler @weissweiler.bsky.social · Mar 31

Tagging our amazing first author @qyao.bsky.social as well

1

Leonie Weissweiler @weissweiler.bsky.social · Mar 31

Yes, hi, thanks for reading 🙂

1

Reposted by Leonie Weissweiler

Leshem (Legend) Choshen @ICML @ACL @lchoshen.bsky.social · Mar 31

Models have preferences like giving inanimate 📦 stuff to animate 👳
Is it that they just saw a lot of such examples in pretraining or is it generalization and deeper understanding?
alphaxiv.org/pdf/2503.20850

2 1 14

Leonie Weissweiler @weissweiler.bsky.social · Mar 27

bsky.app/profile/weis...

Leonie Weissweiler @weissweiler.bsky.social · Nov 26

@kanishka.bsky.social and I have made a starter pack for researchers working broadly on linguistic interpretability and LLMs!

go.bsky.app/F9qzAUn

Please message me or comment on this post if you've noticed someone who we forgot or would like to be added yourself!

3

Leonie Weissweiler @weissweiler.bsky.social · Mar 17

This is really interesting! We did some related work a few years ago: aclanthology.org/2023.acl-lon... and aclanthology.org/2023.finding...

aclanthology.org

1 2

Leonie Weissweiler @weissweiler.bsky.social · Mar 12

✨New paper ✨

RoBERTa knows the difference between "so happy that you're here", "so certain that I'm right" and "so happy that I cried"!

Exciting result (and more) from Josh Rozner along with @coryshain.bsky.social, @kmahowald.bsky.social and myself, go check it out!

1 6

Reposted by Leonie Weissweiler

Siyuan Song✈️COLM @siyuansong.bsky.social · Mar 12

New preprint w/ @jennhu.bsky.social @kmahowald.bsky.social : Can LLMs introspect about their knowledge of language?
Across models and domains, we did not find evidence that LLMs have privileged access to their own predictions. 🧵(1/8)

2 16 60

Leonie Weissweiler @weissweiler.bsky.social · Mar 12

Embracing the PowerPoint suggestion palette here actually 😂

1 1