Discover more about the project and the team here

Today, Paul Lerner @lernerp.bsky.social led our weekly team meeting, presenting his work from #COLING2025 on using LLMs for machine translation of neologisms. Check out the related papers here: 📖 📚

📄 RL for Aligning LLM Agents in Interactive Environments: Quantifying & Mitigating Prompt Overfitting 🔗 by Mohamed Salim Aissi, @ClementRomac, @CartaThomas2, @SLamprier, @pyoudeye, Olivier Sigaud, @LaureSoulier, Nicolas Thome 🤝 With @FlowersINRIA

🎓 It’s happening now! Tristan Luiggi’s PhD defense! 📺 Watch the presentation live on YouTube: 🥂 See you for the post-defense reception in Room 304, Building 65-66! 🎉

🎓 Go see Tristan Luiggi’s PhD defense! 📖 Title: “Leveraging Multi-Dimensional Context & Data Augmentation for NER” 📅 Date: Thursday, February 6, 2025 🕑 Time: 2:00 PM 📍 Location: Amphithéâtre Astier, 4 Place Jussieu, Paris Abstract:

Today, Zineddine TIGHIDET (@Zineddine___t) presented his work on Probing Language Models on Their Knowledge Source, published in the BlackboxNLP workshop at EMNLP! 👏 Check it out here: #EMNLP #BlackboxNLP

Congrats to Nazanin and Guillaume Wisniewski for their work which received the Best Short Paper Award at @coling2025 ! 🥳

@lernerp @yhLiu96 The team in action at #coling2025 ! @_lerner_paul

The first team meeting of 2025 is presented by Arthur Satouf, who has recently joined the team! ✨ He presented his PhD about "Information Retrieval using generative models". His PhD is in collaboration with @airliquidegroup, @UnivParisSaclay, @centralesupelec and the ILLS lab!

(5) "MOSAIC at GENAI Detection Task 3: Zero-Shot Detection Using an Ensemble of Models" by Matthieu Dubois, François Yvon @yvofr.bsky.social, @PiantanidaPablo accepted as poster in the workshop on Detecting AI Generated Content at #COLING2025

(4) "How Transliterations Improve Crosslingual Alignment" by @yhLiu96, Mingyang Wang, Amir Hossein Kargaran, Ayyoob Imani, Orgest Xhelili, Haotian Ye, Chunlan Ma, François Yvon, Hinrich Schütze in collaboration with @LMU_Muenchen ➡️

(3) "Towards the Machine Translation of Scientific Neologisms" by Paul Lerner @lernerp.bsky.social and François Yvon @yvofr.bsky.social accepted as conference paper at #COLING2025 ➡️

(2) "Unlike “Likely”, “Unlike” is Unlikely: BPE-based Segmentation hurts Morphological Derivations in LLMs" by Paul Lerner @lernerp.bsky.social and François Yvon @yvofr.bsky.social accepted as conference paper at #COLING2025 ➡️

(1) "Beyond Surprisal: A Dual Metric Framework for Lexical Skill Acquisition in LLMs" by Nazanin Shafiabadi and Guillaume Wisniewski accepted as conference paper at #COLING2025