
🚀Release!🚀 RoBERTa models pre-trained on millions of tweets. Also fine-tuned models for tweet classification for the #emnlp2020 Findings paper 🎉TweetEval🎉. Models and starter code now online in 🤗@huggingface | huggingface.co/cardiffnlp/tw… #NLProc

Head of Machine Learning at @AMPLYFItech Lecturer at @cardiffuni #nlproc #machinelearning Before: @savanamedica @FulbrightPrgrm @FundlaCaixa Views my own

🚀Release!🚀 RoBERTa models pre-trained on millions of tweets. Also fine-tuned models for tweet classification for the #emnlp2020 Findings paper 🎉TweetEval🎉. Models and starter code now online in 🤗@huggingface | huggingface.co/cardiffnlp/tw… #NLProc

Check out the accepted papers from the Cardiff NLP group and collaborators at upcoming #NLProc and #ML conferences! We'll be presenting several papers at #NAACL2025 (main conference, demos, and workshops), #ICLR2025, and #CVPR2025. Preprints are available for the following papers 👇:

The dates for our 4th Cardiff NLP Workshop have been announced! ✍🏼
Save the date! 👇 4th Cardiff NLP Workshop 📅 14-15 July 2025 📍 Cardiff (Wales, UK) 🔜 More info coming soon!

A few pics showcasing some of the work presented by Cardiff NLP members at #coling2025! 🖼️

Highlights from #WikiWorkshop at #EMNLP2024: Presented our work on Hoax articles (aclanthology.org/2024.wikinl…) and participated in a panel on misinformation with @vlachos_nlp and @IAugenstein . Thanking the organizers and @frimelle for the incredible experience! 1/2

Stereotypes, social biases and topic classification. Three papers from Cardiff NLP presented simultaneously at the same poster session yesterday at #EMNLP2024! 🖼️

The First Workshop on Advancing Natural Language Processing for Wikipedia will happen tomorrow, 16th of November, in conjunction with @emnlpmeeting ! Website: meta.wikimedia.org/wiki/NLP_… Proceedings: aclanthology.org/events/wiki…


Really pleased to say this has been accepted at #EMNLP2024 main
🚨 NEW PAPER ALERT 🚨 Introducing the GlobalBias dataset… We ask Claude 3, GPT 3.5, GPT 4o, and Llama 3 to produce character profiles based on given names from GlobalBias for 40 different gender-by-ethnicity groups. We find that all models displayed stereotypical outputs (1/4)

"WiDe-analysis: Enabling One-click Content Moderation Analysis on Wikipedia’s Articles for Deletion" (Borkakoty and Espinosa-Anke) arxiv.org/pdf/2408.05655


PolygloToxicityPrompts, our multilingual toxicity evaluation benchmark, has been accepted to @COLM_conf and ranked in the top 1% of all submissions! 🥳 Awesome collaboration w/ @kpriyanshu256, @nlpxuhui, @samgehman, @tom_hartvigsen, @MaartenSap; across @UVA @SCSatCMU @allen_ai.
Thrilled to share PolygloToxicityPrompts, a multilingual benchmark to evaluate toxic degenerations in LLMs! Arxiv: arxiv.org/abs/2405.09373 Dataset: huggingface.co/datasets/Toxi… Leaderboard: huggingface.co/spaces/Toxici… w/ @kpriyanshu256 @samgehman @nlpxuhui @tom_hartvigsen @MaartenSap 🧵

🚨 NEW PAPER ALERT 🚨 Introducing the GlobalBias dataset… We ask Claude 3, GPT 3.5, GPT 4o, and Llama 3 to produce character profiles based on given names from GlobalBias for 40 different gender-by-ethnicity groups. We find that all models displayed stereotypical outputs (1/4)

Link to GitHub: github.com/groovychoons/Glob… Link to paper: arxiv.org/abs/2407.06917 If you’d like to discuss anything related to the paper, drop me a DM or comment below. * we refers to myself, @LiamDTurner and @espinosaanke (4/4)

And the 3rd Cardiff NLP Workshop has come to an end... 😢 Hopefully see you all next year, including those who couldn't make it this year! 👋

Day 2 starts in a few hours, let's go! #cardiffnlpworkshop