BathNLP

I'm thrilled to have been awarded the Pivotal Research Fellowship 🩵 I'll be based at the London Initiative for Safe AI (LISA) office next year and will work with Noam Kolt on the compliance dynamics of frontier LLMs, which is a precondition for safer and aligned AI systems.

Huge congratulations to our Lab Head @harish for the promotion to Associate Professor @UniofBath Department of Computer Science! 🥳🎉

Thoroughly enjoyed my first-time poster presentation earlier at EMNLP. My past two primary papers (EMNLP2024 and 2023) were given oral but they were not as fulfilling as this one. Many big plans ahead for UniversalCEFR (universalcefr.github.io). Everyone seems very keen to work on more accessible language resources! 🌍

Pleased to announce our #NLProc shared task @SemEvalWorkshop Task 7: Everyday Knowledge Across Diverse Languages & Cultures. We extend the BLEnD Benchmark to >30 language-culture pairs. [Our task is Junior-friendly, with live Q&A & tutorials.] codabench.org/competitions/1… 1/

Exciting news. Both UniversalCEFR 🇪🇺 and FilBench 🇵🇭 were accepted at #EMNLP2025 Main 🥳 Grateful to my collaborators around the globe for making these projects possible. See you all in Suzhou and Shanghai!

@HaritzPuerto @UKPLab @BathNLP @IGurevych We provide open access to our code, models, data, and results: 📽️Underline: underline.io/events/485/pos… 📄Paper: aclanthology.org/2025.acl-long.… 💻 Code: github.com/UKPLab/acl2025… 🤗 Models: huggingface.co/collections/ha… 📂 Data: tudatalib.ulb.tu-darmstadt.de/handle/tudatal… 🌐 Website: ukplab.github.io/acl2025-divers… (9/🧵)

At first I was not sure🤔, but on second thought, I knew what to do!!!💡😃 📢 Diverse Chains of Thought help LLMs refine their Reasoning!! @haritzpuerto will be presenting our work at #ACL2025NLP 🇦🇹 on Wednesday 30th at 11:00 #NLProc A 🧵👇

We are excited to share that two long main conference papers from our lab have been accepted at ACL 2025 in Vienna! 🎉 We are incredibly grateful to collaborate with such an impressive group of researchers worldwide @UKPLab @seacrowd_ai 🤗 @harish @josephimperial_

The trial data has just been released to registered participants. There’s still time for your team to join! #emnlp2025 #nlproc

Join our small but highly productive lab! 👇

The CardiffNLP Workshop 🏴󠁧󠁢󠁷󠁬󠁳󠁿 is one of my favorite NLP events in the UK. Love the selection of topics this year, my favorites being Rethinking LLM Evals @mziizm, Responsible AI @c_christodoulop and Steering Vectors @TheZaraSiddique. Amazingly organized as always @Cardiff_NLP @nedjmaou @feralvam @CamachoCollados et al. 🤗

🎉 The Cardiff NLP Workshop kicks off this Monday (14 July)! Check out the full schedule on our website 👉 cardiffnlpworkshop.org/schedule We’ve got an exciting lineup: ✨ Talks by fantastic speakers 💻 A tutorial on steering vectors 🖼️ Poster session and networking opportunities 🔄 An interdisciplinary session 💬 A panel discussion Looking forward to welcoming you to Cardiff!

📢 Call for Participation: TSAR 2025 Shared Task 📢 Simplify English paragraphs to a specified CEFR level 🔹 No training data 🔹 Eval: CEFR match, meaning preservation, ref similarity 🗓️ Starts July 16 🔗 Info + registration: tsar-workshop.github.io/shared-task #EMNLP2025 #TextSimplification

🚨 New global collaboration & dataset paper! UniversalCEFR: Enabling Open Multilingual Research on Language Proficiency Assessment 🌍 We introduce UniversalCEFR, an initiative to build a growing, open, multilingual, and multidimensional resource for CEFR-based language proficiency assessment. 🌐 Project Website: universalcefr.github.io 📄 Paper: arxiv.org/abs/2506.01419 📀 HF Org & Data: huggingface.co/UniversalCEFR 📀 Github: github.com/UniversalCEFR

Is In-Context Learning (ICL) in LLMs Memorisation? Emergence? Some Algorithmic Capability? 🤔 📢New work exploring ICL in LLMs: arxiv.org/abs/2505.11004 💡Key Finding: ICL capabilities are linked to token frequency 🤨 Strap in for the unexpected 🤯 A 🧵👇 #NLProc #LLMs

This has been accepted to ACL2025 (Main) 🥳 Happy to have contributed so much to this project, along with my amazing collaborators at @seacrowd_ai !

Great to hear from @josephimperial_, an @RealAAAI / @acm_sigai doctoral consortium participant and PhD student at @UniofBath. He told us about his research aligning generative AI with technical standards aihub.org/2025/04/02/int…

In London attending the AI Standards Hub Global Summit organized by @turinginst. I’m also excited to share that I joined the British Standards Institution (@BSI_UK) 🇬🇧—the UK’s national standards body—as a committee member for the AI working group ART/1 and will work with CEN-CENELEC’s Joint Task Force 21. This is the working group in-charge of developing standards for the EU AI Act.

One of the largest NLP meetups in Cardiff and UK! Open to all NLP researchers 📢

I'll be attending #AAAI2025 Doctoral Consortium in Philadelphia 🇺🇸 next week to present my current PhD research on controlling NLG for standard alignment. Happy to chat or meet up! I don't know what Philly is like so if you have recommendations (food or places) I'd love to hear them!🤗 Also, check out some of my work so far: 1. 🆕Standardizing Intelligence: Aligning Generative AI for Regulatory and Operational Compliance (hal.science/hal-04936214) 2. Standardize: Aligning Language Models with Expert-Defined Standards for Content Generation (EMNLP Main, arxiv.org/abs/2402.12593) 3. Flesch or Fumble? Evaluating Readability Standard Alignment of Instruction-Tuned Language Models (EMNLP Findings, arxiv.org/abs/2407.13297) 4. Flesch or Fumble? Evaluating Readability Standard Alignment of Instruction-Tuned Language Models (GEM Workshop, arxiv.org/abs/2309.05454)