Lightnews — Scholar-powered news

Flor Plaza @florplaza.bsky.social · 2d

📢 Are you interested in a PhD in #NLProc to study and improve how AI model emotions and social signals?

🚨Exciting news:🚨 I’m hiring a PhD candidate at LIACS,
@unileiden.bsky.social.

📍 Leiden, The Netherlands
📅 Deadline: 17 Nov 2025

👉 Position details and application link: tinyurl.com/5x5v6zsa

PhD Candidate in Emotionally and Socially Aware Natural Language Processing

The Faculty of Science and the Leiden Institute of Advanced Computer Science (LIACS) are looking for a:PhD Candidate in Emotionally and Socially Aware Natural Language Processing (1.0fte)Project descr...

tinyurl.com

8 9

Reposted by Flor Plaza

Dirk Hovy @dirkhovy.bsky.social · 26d

Comprehensive preprint out now, check it out (feedback welcome!).

Joachim Baumann @joachimbaumann.bsky.social · 26d

🚨 New paper alert 🚨 Using LLMs as data annotators, you can produce any scientific result you want. We call this **LLM Hacking**.

Paper: arxiv.org/pdf/2509.08825

$We present our new preprint titled "Large Language Model Hacking: Quantifying the Hidden Risks of Using LLMs for Text Annotation". We quantify LLM hacking risk through systematic replication of 37 diverse computational social science annotation tasks. For these tasks, we use a combined set of 2,361 realistic hypotheses that researchers might test using these annotations. Then, we collect 13 million LLM annotations across plausible LLM configurations. These annotations feed into 1.4 million regressions testing the hypotheses. For a hypothesis with no true effect (ground truth $p > 0.05$), different LLM configurations yield conflicting conclusions. Checkmarks indicate correct statistical conclusions matching ground truth; crosses indicate LLM hacking -- incorrect conclusions due to annotation errors. Across all experiments, LLM hacking occurs in 31-50\% of cases even with highly capable models. Since minor configuration changes can flip scientific conclusions, from correct to incorrect, LLM hacking can be exploited to present anything as statistically significant.$

1 3

Reposted by Flor Plaza

Joachim Baumann @joachimbaumann.bsky.social · 26d

🚨 New paper alert 🚨 Using LLMs as data annotators, you can produce any scientific result you want. We call this **LLM Hacking**.

Paper: arxiv.org/pdf/2509.08825

$We present our new preprint titled "Large Language Model Hacking: Quantifying the Hidden Risks of Using LLMs for Text Annotation". We quantify LLM hacking risk through systematic replication of 37 diverse computational social science annotation tasks. For these tasks, we use a combined set of 2,361 realistic hypotheses that researchers might test using these annotations. Then, we collect 13 million LLM annotations across plausible LLM configurations. These annotations feed into 1.4 million regressions testing the hypotheses. For a hypothesis with no true effect (ground truth $p > 0.05$), different LLM configurations yield conflicting conclusions. Checkmarks indicate correct statistical conclusions matching ground truth; crosses indicate LLM hacking -- incorrect conclusions due to annotation errors. Across all experiments, LLM hacking occurs in 31-50\% of cases even with highly capable models. Since minor configuration changes can flip scientific conclusions, from correct to incorrect, LLM hacking can be exploited to present anything as statistically significant.$

5 94 260

Reposted by Flor Plaza

Workshop on Online Abuse and Harms @woahworkshop.bsky.social · Aug 25

Excited for WOAH’s 10th anniversary? 😍

We're launching an open call for new organisers!
Our goal: diversify the team and bring in fresh perspectives.

🗓️ Apply by September 12
🔗 forms.gle/aiFs35vwDXnt...

4 3

Reposted by Flor Plaza

MilaNLP Lab @milanlp.bsky.social · Jul 31

#TBT #NLProc Plaza-del-Arco et al.'s 'Divine LLaMAs..' (2024) addresses stereotypes on religion in Large Language Models and their social impact. Tech for all! #DigitalEquity

aclanthology.org

3 6

Reposted by Flor Plaza

Tommaso Caselli @tommasoc80.bsky.social · Jul 27

🗓️ Fri 01/08, I’ll be presenting two papers at # WOAH2025

“Learning from Disagreement: Entropy-Guided Few-Shot Selection for Toxic Language Detection” with @florplaza.bsky.social
💻 github.com/tommasoc80/w...
📝 aclanthology.org/2025.woah-1....

GitHub - tommasoc80/woah_2025_shot_selection

Contribute to tommasoc80/woah_2025_shot_selection development by creating an account on GitHub.

github.com

1 1 4

Reposted by Flor Plaza

Debora Nozza @deboranozza.bsky.social · Jul 27

"Blue-haired, misandriche, rabiata: Tracing the Connotation of 'Feminist(s)' Across Time, Languages and Domains"
With @arimuti.bsky.social, @saragemelli.bsky.social, Emanuele Moscato, Emilie Francis, Amanda Cercas Curry, @florplaza.bsky.social #WOAH2025

1 2 3

Reposted by Flor Plaza

Maria Antoniak @mariaa.bsky.social · Jul 17

The #ACL2025 #ACL2025NLP feed is up and running! It matches both hashtags and any posts from or mentions of @aclmeeting.bsky.social

Pin it to your home 📌 and enjoy!

bsky.app/profile/did:...

2 14 48

Reposted by Flor Plaza

AILC-NLP @ailc-nlp.bsky.social · Jun 9

Stiamo vivendo una rivoluzione tecnologica, ma quanto conosciamo realmente l'uso dell'IA nella nostra quotidianità? Aiuta a scoprirlo compilando questo breve questionario (10 min): bit.ly/sondaggio_ai...

7 8

Reposted by Flor Plaza

Giuseppe Attanasio @gattanasio.cc · Jun 16

(totally unbiased opinion): Cool paper!

MilaNLP Lab @milanlp.bsky.social · Jun 16

#MemoryMonday #NLProc 'Interpretability Informs Gender Bias Mitigation in Machine Translation'- Attanasio et al. (2023). Their new research tackles gender bias in machine translations using a simple, few-shot learning method for fairness.

A Tale of Pronouns: Interpretability Informs Gender Bias Mitigation for Fairer Instruction-Tuned Machine Translation

Giuseppe Attanasio, Flor Miriam Plaza del Arco, Debora Nozza, Anne Lauscher. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.

aclanthology.org

4 7

Reposted by Flor Plaza

MilaNLP Lab @milanlp.bsky.social · Jul 16

🎉 The @milanlp.bsky.social lab is excited to present 15 papers and 1 tutorial at #ACL2025 & workshops! Grateful to all our amazing collaborators, see everyone in Vienna! 🚀

8 11

Reposted by Flor Plaza

Workshop on Online Abuse and Harms @woahworkshop.bsky.social · Jul 15

🥁 #WOAH2025 Speakers🥁

We are excited to announce an amazing panel of 3 speakers who will discuss our theme "Harms Beyond Hate Speech"

🗣️ @cordeliamoore.bsky.social
🗣️ @fvancesco.bsky.social
🗣️ @katejsim.bsky.social

Full details: www.workshopononlineabuse.com/keynotes.html

🤩 See you on August 1!

1 6 13

Reposted by Flor Plaza

Aswathy Velutharambath @aswathyve.bsky.social · Jun 25

✨ The First NLPSI Workshop @ ICWSM 2025 is a wrap!
Huge thanks to the speaker (Bennett Kleinberg), authors, reviewers, attendees, and co-organizers (Sofie Labat, Neele Falk, @florplaza.bsky.social, @romanklinger.de & Veronique Hoste) who made it a success.

nlpsi-workshop.github.io/highlights/

2 4

Reposted by Flor Plaza

Dirk Hovy @dirkhovy.bsky.social · May 23

Had the chance to revamp a PhD class on CSS. We're now using Colab notebooks to use Python for data analysis.

Check out the new syllabus (w/ reading, slides, background notebooks, and exercises plus solutions).

📚👇 bit.ly/438R6Hk

#ComputationalSocialScience #PhD

Syllabus

Syllabus: Computational Methods for Social and Political Science, Spring 2025 Recommended Reading: Christian, B., & Griffiths, T. (2016). Algorithms to live by: The computer science of human decisio...

docs.google.com

2 9

Reposted by Flor Plaza

Workshop on Online Abuse and Harms @woahworkshop.bsky.social · May 23

📢 #WOAH2025 invites non-archival submissions of relevant papers accepted to the ✨Findings of @aclmeeting.bsky.social✨

To present your Findings paper at #WOAH2025, complete the form by 📆June 10, 2025.

📝Form: docs.google.com/forms/d/e/1F...

docs.google.com

4 5

Flor Plaza @florplaza.bsky.social · Apr 24

It was a pleasure to speak at the #SSaLM workshop at the Weizenbaum Institute, where I gave my invited talk, "The Heart of Language Models: Decoding Emotion Biases and Societal Stereotypes."

Thank you, Fatma Elsafoury and Jan Batzner, for inviting me and for the fantastic organization!

1

Reposted by Flor Plaza

Debora Nozza @deboranozza.bsky.social · Apr 7

👥 Join us for the 1st Workshop on Protecting Women Online at TheWebConf!
📅 April 28, 2025 | 📍Sydney, Australia
🌐 Program & details: tsww25.github.io

Featuring amazing invited talks & discussions on tackling #VAWG — don’t miss it!

#TheWebConf25 #NLProc

3 3

Flor Plaza @florplaza.bsky.social · Mar 31

Don't miss the #WOAH2025 deadline!

Workshop on Online Abuse and Harms @woahworkshop.bsky.social · Mar 31

📆 Reminder! 📆

🚀 The #WOAH2025 submission deadline is just over two weeks away! (April 18, 2025, Anywhere on Earth!)

🔗 CfP: workshopononlineabuse.com/cfp.html

We're excited to see your submissions! 🤩

#ACL2025 #NLProc

Call For Papers

The 9th Workshop on Online Abuse and Harms (WOAH) at ACL 2025.

workshopononlineabuse.com

1 1

Reposted by Flor Plaza

Debora Nozza @deboranozza.bsky.social · Mar 28

This week, I had the pleasure of giving an invited talk at the #HumanCLAIM workshop in Göttingen on hate speech and pluralism in subjective NLP 🗣️✨ Big thanks to Lisa Beinborn for the invitation and the organization! #NLProc

1 10

Reposted by Flor Plaza

Suzan Verberne @suzanv.bsky.social · Mar 28

3 days left to respond to this PhD vacancy with Niki van Stein:

PhD Candidate in Explainable NLP for Regulatory Compliance Detection www.universiteitleiden.nl/en/vacancies...

Part of the ICAI lab AI4Oversight, collaboration with TNO and 4 inspectorates, together with 5 other PhD students

PhD Candidate in Explainable NLP for Regulatory Compliance Detection

The Faculty of Science and the Leiden Institute of Advanced Computer Science (LIACS) are looking for a:PhD Candidate in Explainable NLP for Regulatory Compliance Detection Are you passionate about Exp...

www.universiteitleiden.nl

1 2

Flor Plaza @florplaza.bsky.social · Mar 27

🗣️ Esta tarde a las 🕔 17:00 (CEST), tendré el placer de hablar sobre ✨"Análisis de emociones en PLN: Desafíos, Aplicaciones y Consideraciones Éticas"✨en las Charlas de la SEPLN. ¿Te gustaría unirte?

➡ Regístrate aquí: www.sepln.org/actualidad/e... (online)

C-SEPLN: Análisis de emociones en PLN: Desafíos, Aplicaciones y Consideraciones Éticas | Sociedad Española de Procesamiento del Lenguaje Natural

www.sepln.org

Flor Plaza @florplaza.bsky.social · Mar 25

💙💙💙💙

Reposted by Flor Plaza

MilaNLP Lab @milanlp.bsky.social · Mar 25

Thrilled for Flor and her new position! 🌟 Congrats on this exciting chapter—you're going to be amazing!

Flor Plaza @florplaza.bsky.social · Mar 24

New chapter ahead! 🚀

I’ve joined as an Assistant Professor at LIACS, Leiden University @unileiden.bsky.social! 🇳🇱🌷

The past two years at @milanlp.bsky.social have been incredibly rewarding and fun. Thanks to the fantastic MilaNLP crew for making it unforgettable!

1 3 6

Flor Plaza @florplaza.bsky.social · Mar 24

New chapter ahead! 🚀

I’ve joined as an Assistant Professor at LIACS, Leiden University @unileiden.bsky.social! 🇳🇱🌷

The past two years at @milanlp.bsky.social have been incredibly rewarding and fun. Thanks to the fantastic MilaNLP crew for making it unforgettable!

5 24

Flor Plaza @florplaza.bsky.social · Mar 5

Find more information here: nlpsi-workshop.github.io

This link will take you to a page that’s not on LinkedIn

lnkd.in

2