Debora Nozza

Thanks Clara Meister for presenting "Parity-Aware Byte-Pair Encoding: Improving Cross-lingual Fairness in Tokenization" at our lab seminar.

#NLProc #tokenization #fairness

📚 For today’s reading group @arimuti.bsky.social presented Emergent Misalignment: Narrow Finetuning Can Produce Broadly Misaligned LLMs (Betley et al., 2025).

🧩 arxiv.org/abs/2502.17424

#NLProc #AIAlignment #LLMs

Explaining Speech Classification Models via Word-Level Audio Segments and Paralinguistic Features Eliana Pastor, Alkis Koudounas, Giuseppe Attanasio, Dirk Hovy, Elena Baralis. Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long...

#TBT #NLProc "Explaining Speech Classification Models" by Pastor et al. (2024) makes speech classification more transparent! 🔍 Their research reveals which words matter most and how tone and background noise impact decisions.

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...

📢 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

Measuring Harmful Representations in Scandinavian Language Models Samia Touileb, Debora Nozza. Proceedings of the Fifth Workshop on Natural Language Processing and Computational Social Science (NLP+CSS). 2022.

#MemoryModay #NLProc 'Measuring Harmful Representations in Scandinavian Language Models' uncovers gender bias, challenging Scandinavia's equity image. #MachineLearning

Wisdom of Instruction-Tuned Language Model Crowds. Exploring Model Label Variation Flor Miriam Plaza-del-Arco, Debora Nozza, Dirk Hovy. Proceedings of the 3rd Workshop on Perspectivist Approaches to NLP (NLPerspectives) @ LREC-COLING 2024. 2024.

#TBT #NLProc Explore 'Wisdom of Instruction-Tuned LLM Crowds' by Plaza et al. LLM labels outperform single models in tasks & languages. But few-shot can't top zero-shot. Supervised models rule.

🚨 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.

📣 New Preprint!
Have you ever wondered what the political content in LLM's training data is? What are the political opinions expressed? What is the proportion of left- vs right-leaning documents in the pre- and post-training data? Do they correlate with the political biases reflected in models?

📖For our last Reading Group @donyarn.bsky.social presented
"Culture is Everywhere: A Call for Intentionally Cultural Evaluation" by Oh et al.

Paper: arxiv.org/pdf/2509.01301

#NLProc

Universal Joy A Data Set and Results for Classifying Emotions Across Languages Sotiris Lamprinidis, Federico Bianchi, Daniel Hardt, Dirk Hovy. Proceedings of the Eleventh Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis. 2021.

#MemoryModay #NLProc 'Universal Joy: A Data Set and Results for Classifying Emotions Across Languages' by Lamprinidis et al. (2021) explores how AI research affects our planet. Tech can be green too! #SustainableTech

MilaNLPers at CLiC-it 2025 presenting "Probing Feminist Representations: A Study of Bias in LLMs and Word Embeddings"

Check the paper at clic2025.unica.it/wp-content/u...

#NLProc #clicit25

What makes LLMs agree even when they shouldn’t? 🤔
At our last seminar’s lab, Jan Batzner presented The Brief History of LLM Sycophancy as We Know It.

#NLProc #sycophancy #LLM

#TBT #NLProc 'Classist Tools: Social Class Correlates with Performance in NLP' by Curry et al. (2024) explores AI's hidden energy problem, and how machine learning impacts environmental sustainability. Tech can be green! #CleanTech

Benchmarking Post-Hoc Interpretability Approaches for Transformer-based Misogyny Detection Giuseppe Attanasio, Debora Nozza, Eliana Pastor, Dirk Hovy. Proceedings of NLP Power! The First Workshop on Efficient Benchmarking in NLP. 2022.

#MemoryModay #NLProc 'Benchmarking Post-Hoc Interpretability Approaches for Transformer-based Misogyny Detection' - Attanasio et al. Explores reliability of interpretability in hate speech detection.

📖For this Thursday's Reading Group @elisabassignana.bsky.social presented "The sociolinguistic foundations of language modeling" by Grieve et al.

Paper: www.frontiersin.org/journals/art...

#NLProc #LLM #sociolinguistics

Geolocation with Attention-Based Multitask Learning Models Tommaso Fornaciari, Dirk Hovy. Proceedings of the 5th Workshop on Noisy User-generated Text (W-NUT 2019). 2019.

#TBT #NLProc 'Geolocation with Attention-Based Multitask Learning Models' by Tommaso Fornaciari, @dirkhovy.bsky.social (2019) reveals how online political talks can become one-sided. Breaking out of our bubbles! #SocialMedia

🎓We're back with Fridays' lab seminars! Today we had
Suyash Fulay presenting "Truth, Political Bias, and AI Representation".

#NLProc

Dense Node Representation for Geolocation Tommaso Fornaciari, Dirk Hovy. Proceedings of the 5th Workshop on Noisy User-generated Text (W-NUT 2019). 2019.

#MemoryModay #NLProc 'Dense Node Representation for Geolocation' by Fornaciari & @dirkhovy.bsky.social reveals efficient geolocation methods using node2vec & doc2vec models. Greater network size, less parameters. /publication/2019_m2v/2019_m2v

📖For this Thursday's Reading Group @deboranozza.bsky.social
presented two papers on #sycophancy in #LLMs.

Papers: arxiv.org/pdf/2505.13995, arxiv.org/pdf/2310.13548

#NLProc

#TBT #NLProc 'MilaNLP @ WASSA: Does BERT Feel Sad When You Cry?' by Fornaciari et al. (2021) indicates that emotion and empathy are not related tasks for prediction.

#MemoryModay #NLProc 'My Answer is C' by Wang et al. (2024) underscores the scrutiny needed for full text responses in LLMs multi-choice evaluations.

We're back with the reading group! 🚀
Today @paul-rottger.bsky.social presented The Levers of Political Persuasion with Conversational AI by Hackenburg et al. (2025)

Paper: arxiv.org/pdf/2507.13919

#NLProc

#TBT #NLProc 'Temporal and Second Language Influence on Intra-Annotator Agreement and Stability in Hate Speech Labelling' Abercrombie, @dirkhovy.bsky.social, & Prabhakaran (2023). NLP models: inclusive or not?

#MemoryModay #NLProc Hung et al.'s 2023 paper, 'Can Demographic Factors Improve Text Classification?' finds demographic adaptations of Transformer NLP models don't notably boost performance.

#TBT #NLProc Outstanding Paper at ACL 2024! 'Towards More Meaningful Evaluations for Values and Opinions in Large Language Models' by @paul-rottger.bsky.social et al. Enhancing political bias evaluation.

#MemoryModay #NLProc 'Detecting Misogynous Memes with Text & Image Modalities' by Attanasio, @deboranozza.bsky.social, Bianchi. Their novel system uses Perceiver IO, surpassing all previous benchmarks.

#TBT #NLProc 'XLM-EMO: Multilingual Emotion Prediction in Social Media Text' by Bianchi, @deboranozza.bsky.social, @dirkhovy.bsky.social (2022) advances cross-language emotion detection, especially for low-resource languages.

#MemoryModay #NLProc Plaza-del-Arco, @debora_nozza, @dirkhovy.bsky.social's 2024 paper "Wisdom Instruction-Tuned Language Model Crowds" shows that multiple LLMs can be BETTER than a single model & specialize across tasks & languages.

#TBT #NLProc 'Make Natural Language Processing About People Again' by @dirkhovy.bsky.social (2018) uncovers how AI models portray different religions and emotions. #AIEthics

#MemoryModay #NLProc #TBT #NLPproc 'Pipelines for Social Bias Testing of Large Language Models' by @deboranozza.bsky.social, Federico Bianchi, @dirkhovy.bsky.social (2022). Proposes social bias tests akin to software testing in AI dev pipelines.