TrustLLM
@TrustLLM_
Followers
39
Following
8
Media
12
Statuses
48
TrustLLM is a research project; democratising trustworthy and factual large language model technology for Europe. TrustLLM is funded by Horizon Europe.
Joined March 2024
📢 Stay updated on our research, publications, events and more by subscribing to our newsletter! Sign up here: #trustworthyAI #socialAI #machinelearning #artificialintelligence #ai #generativeai #llm #llms.
0
1
2
It was great to see the TrustLLM team come together in Reykjavík! 🇮🇸 We're proud to collaborate with partners like @fzj_jsc as we work toward a multilingual, trustworthy, and energy-efficient AI ecosystem for Europe.
🚀JSC joined the @TrustLLM_ meeting in Reykjavík (June 11–13), advancing Europe's mission for multilingual, trustworthy AI. As lead of WP6, we’re building the HPC foundations for scalable, ethical LLMs. Proud to help shape AI that reflects Europe's diversity. 🌍
0
0
0
📣 Join us on Iceland for a session on the future of LLMs in Europe! 📷 11 June, 9:00–12:00 📷 Keynotes incl. Lilja Dögg Alfreðsdóttir, Lilja Dögg Jónsdóttir & Edouard Geoffrois. Read more:
0
0
0
📣 Join us for a session on the future of LLMs in Europe!.🗓️ 11 June, 9:00–12:00.📍 Veröld - Vigdís’ House, University of Iceland.🔊 Keynotes incl. Lilja Dögg Alfreðsdóttir @liljaalfreds, Lilja Dögg Jónsdóttir & Edouard Geoffrois.Read more👉
trustllm.eu
Facebook X LinkedIn
0
0
0
Join us for a half-day seminar exploring how ethics and trust can be built into large language models. Details: #trustworthyAI #socialAI #machinelearning #ai #generativeai #llm #llms #artificialintelligence
0
0
0
📖 New insights from "Encoder vs Decoder" reveal that encoder models outperform decoders on multilingual NLU tasks—even with fewer parameters. Read more:
arxiv.org
This paper explores the performance of encoder and decoder language models on multilingual Natural Language Understanding (NLU) tasks, with a broad focus on Germanic languages. Building upon the...
0
0
1
Our consortium brings together unique expertise in developing LLMs alongside leading NLP researchers and organisations focused on technology transfer. Explore our interviews with some of them here:
0
1
1
Danila Petrelli recently showcased her data management expertise in the TrustLLM project and other initiatives at AI Labs. Thanks to Danila & AI Sweden! #trustworthyAI #socialAI #machinelearning #artificialintelligence #ai #generativeai #llm #llms"
0
1
2
🌟 Meet Annika Simonsen, PhD student at the University of Iceland, advancing Germanic language models in the TrustLLM project! Read the interview: #TrustworthyAI #SocialAI #MachineLearning #AI #GenerativeAI #LLM #LLMs #LanguageModels #GermanicLanguages
0
0
1
Proposal submitted by Michael Fromm @effi288, Mehdi Ali @MehdiAli2307, and Fredrik Heintz @FredrikHeintz.
0
0
2
TrustLLM secures 500K node hours on EuroHPC Leonardo BOOSTER for AI Act-compliant LLM training! 👉 Read more: #TrustLLM #EuroHPC #trustworthyAI
1
2
3
🚀 @AISweden: Driving data excellence in TrustLLM! Magnus Sahlgren shares insights on data governance, data collection & curation to support the development of European LLMs. Read more: #AI #TrustLLM #AISweden #MagnusSahlgren
0
0
1
📖 This week, we are highlighting the publication "Train More Parameters But Mind Their Placement: Insights into Language Adaptation with PEFT." Jenny Kunz explores enhancing LLMs with unstructured text corpora, focusing on Icelandic. Read more:
arxiv.org
Smaller LLMs still face significant challenges even in medium-resourced languages, particularly when it comes to language-specific knowledge -- a problem not easily resolved with...
0
0
1
📢 Stay updated on our research, publications, events and more by subscribing to our newsletter! Sign up here: #trustworthyAI #socialAI #machinelearning #artificialintelligence #ai #generativeai #llm #llms.
0
0
0
📖 Discover insights from "How Reliable Are Automatic Evaluation Methods for Instruction-Tuned LLMs?". This study explores the reliability of automatic evaluation methods compared to human evaluations across a broad range of tasks. Read more:
aclanthology.org
Ehsan Doostmohammadi, Oskar Holmström, Marco Kuhlmann. Findings of the Association for Computational Linguistics: EMNLP 2024. 2024.
0
0
0
📖 Insights from "Tokenizer Choice For LLM Training: Negligible or Crucial?" reveal that tokenizer choice impacts LLM performance and costs. Multilingual tokenizers need larger vocabularies, leading to performance drops and higher costs. Read the paper:
aclanthology.org
Mehdi Ali, Michael Fromm, Klaudia Thellmann, Richard Rutmann, Max Lübbering, Johannes Leveling, Katrin Klug, Jan Ebert, Niclas Doll, Jasper Buschhoff, Charvi Jain, Alexander Weber, Lena Jurkschat,...
0
0
0
📖 New insights from "Do Multilingual Large Language Models Mitigate Stereotype Bias?" reveal that multilingual LLMs reduce bias and improve accuracy over monolingual models. Read more:
aclanthology.org
Shangrui Nie, Michael Fromm, Charles Welch, Rebekka Görge, Akbar Karimi, Joan Plepi, Nazia Mowmita, Nicolas Flores-Herr, Mehdi Ali, Lucie Flek. Proceedings of the 2nd Workshop on Cross-Cultural...
0
0
0
🌟 Meet our project partner: @NTNU! 🌟 As part of TrustLLM, NTNU is key in developing uniform evaluation frameworks for Germanic language modelling tasks, creating Norwegian datasets, and providing industry use cases. They lead @SFI_NorwAI, innovating sustainable AI solutions.
0
1
3
🌟 Meet Benjamin Kille from @NTNU! 🌟 As part of TrustLLM, Benjamin shares insights on AI trustworthiness, the project's impact, and the challenge of objective evaluation. "Evaluating AI models objectively remains a key challenge," he notes. Read more:
0
0
2
🌟 Join us for the NB-REAL Workshop 2025! 🌟 Welcome to our first workshop on responsible evaluation and alignment of LLMs for Nordic and Baltic languages. 📅 2 March 2025 📍 Tallinn, Estonia. Hybrid format available! 👉 Read more: #trustworthyAI #LLM
0
0
0
📖 New insights from "The Impact of Language Adapters in Cross-Lingual Transfer for NLU"! Results show varying impacts across tasks, languages and models. Read more:
aclanthology.org
Jenny Kunz, Oskar Holmström. Proceedings of the 1st Workshop on Modular and Open Multilingual NLP (MOOMIN 2024). 2024.
0
0
0