@grinbergnir

Assistant prof. @BenGurionU, dad^3, train+bike, tired/aspiring barista. Studying how information systems can better serve us using ML, NLP, Stats, CSS, HCI.

 Tel-Aviv, Israel
nirg.net
Joined January 2010
  • Tweets494
  • Following392
  • Followers672
58Media
🚨New paper🚨out in @SocMedSoc w/ John Wagner and @jennifer_oser 🎉 We all know who's a political influencer on social media, right? Well... we may THINK we do, but what we're actually measuring is very different. How different? 🧵👇
2
4
14
451
Sorry, the correct tag should have been @SocialMedia_Soc
18
The overarching story here is that to build knowledge about social media influencers, we need clearer, multidimensional frameworks (rather than a single metric), and we need to be more transparent about what we actually measure.
1
10
Read the full open-access article here: doi.org/10.1177/205630512614…. We are grateful for comments received from participants of the Digital Democracy Workshop at UZH, MPSA, and ICA conferences, and ERC funding.

How Definitions Shape Our Understanding of Political Influencers on Social Media - Jennifer Oser,...

Political discourse now circulates through sprawling social media networks where multiple actors wield influence in addition to traditional elites. Yet scholarl...

journals.sagepub.com
1
12
Nir Grinberg retweeted
Raghav Singhal
@ragghhavvvAug 14
New paper! Synthetic Persona Pretraining (SPP): Alignment from Token Zero Rather than aligning models after pretraining, we do so from the very start. SPP models are more faithful to our constitution, less misaligned, & more robust to jailbreaks. Advantage grows with scale! 🧵
3
24
8
111
15,135
At #ACL2026 in San Diego? 🌴 Come see our poster on JudgeMeNot, on how to personalize LLMs to emulate judicial reasoning. 📍 Grand Hall, Poster Session F 🕘 Tomorrow (Tue, July 7), 09:00 - 10:30 AM
Nir Grinberg@grinbergnirMay 7
🚨 Excited to share our new paper to appear in the Findings of ACL 2026! JudgeMeNot asks: Can LLMs learn the specific reasoning signature of an individual judge? We tackle this in low-resource settings of Hebrew case law. 🔗Paper: arxiv.org/abs/2604.18041 🧵on nicer platforms👇
The illustrative diagram shows our approach for training personalized LLMs to emulate Judge A (textualist) vs. Judge B (purposivist) and how those trained models might respond to a legal query.

ALT The illustrative diagram shows our approach for training personalized LLMs to emulate Judge A (textualist) vs. Judge B (purposivist) and how those trained models might respond to a legal query.

4
435
1/ Our new book with Zhen Guo, Nir Grinberg and David Lazer is out in @CambridgeUP Elements: Who Follows Whom? The Role of Geography and Similarity in Online Attention Networks cambridge.org/core/elements/…

Who Follows Whom?

Cambridge Core - Politics: General Interest - Who Follows Whom?

cambridge.org
1
1
2
148
🚨 Excited to share our new paper to appear in the Findings of ACL 2026! JudgeMeNot asks: Can LLMs learn the specific reasoning signature of an individual judge? We tackle this in low-resource settings of Hebrew case law. 🔗Paper: arxiv.org/abs/2604.18041 🧵on nicer platforms👇
The illustrative diagram shows our approach for training personalized LLMs to emulate Judge A (textualist) vs. Judge B (purposivist) and how those trained models might respond to a legal query.

ALT The illustrative diagram shows our approach for training personalized LLMs to emulate Judge A (textualist) vs. Judge B (purposivist) and how those trained models might respond to a legal query.

2
1
7
498
more replies
LinkedIn: linkedin.com/posts/nir-grinb…

#nlp #acl2026 #legaltech | Nir Grinberg

🚨 Excited to share our new paper to appear in the Findings of ACL 2026! "JudgeMeNot" asks: Can LLMs learn the specific reasoning signature of an individual judge? We tackle this in the low-resource...

linkedin.com
1
55
🎉 Super proud of this work out of our lab at @BenGurionUni ! Huge thanks to my co-authors Itay Razumenko and Arnon Sturm.
48
Can social media detect economic shocks before official data does? A new PNAS Nexus study led by @grinbergnir and @spfraib shows that AI models tracking job-loss disclosures on social media can predict U.S. unemployment insurance claims up to two weeks early, often
2
1
8
236
New paper out in @PNASNexus We show how skewed social media data can still be used to reliably estimate unemployment, not just nationally but down to the city level. 📈 doi.org/10.1093/pnasnexus/pg… 1/N

Can social media reliably estimate unemployment?

Abstract. Digital trace data hold tremendous potential for measuring policy-relevant outcomes in real-time, yet its reliability is often questioned. Here,

academic.oup.com
1
2
2
151
Nir Grinberg retweeted
Kevin Yang@yang3kc21 Jul 2025
ICYMI, our DomainDemo dataset, which describes how different demographic groups share domains on Twitter, is now available to download! 📄 Data descriptor: doi.org/10.1038/s41597-025-0… 📈 Interactive app to explore the data: domaindemo.info 💽 Dataset: doi.org/10.5281/zenodo.15151…
7
12
1,594
A belated 🧵 about our latest #KDD2024 paper. We developed a novel machine-learning approach for identifying fake news *sources* by leveraging network info and active learning 👇 doi.org/10.1145/3637528.3671…

Leveraging Exposure Networks for Detecting Fake News Sources | Proceedings of the 30th ACM SIGKDD...

dl.acm.org
1
6
1
20
1,649
more replies
This work wouldn't have been possible without the amazing work of Maor Reuben, @Ldfriedl & @PPuzis. Some of the intuition for looking at this problem came from past work done w/ @_kenny_joseph and @davidlazer.
1
149
Also check out the @BenGurionU news story about the study, which reminded me to write this 🧵🤦bgu.ac.il/en/news-and-articl…

New machine learning model can identify fake news sources more reliably

BGU researchers have developed a method to help fact-checkers keep up with the increasing volumes of misinformation on social media.

bgu.ac.il
110
Load more