Sign in

Julen Etxaniz

@juletxara.bsky.social
70 followers 184 following 0 posts

PhD Student in Language Analysis and Processing at HiTZ Zentroa - IXA UPV/EHU Working on Improving Language Models for Low-resource Languages.

PostsRepliesMedia
Reposted by Julen Etxaniz
Aholab Taldea @aholab.bsky.social · 19/12/2025
Aurtengo LibriBrain Competition-ean parte hartu dugu, eta gure neural2speech sistema lehen postua lortu du Phoneme Classification Standard Track-ean. Aurrerapausu galanta! neural-processing-lab.github.io/2025-libribr...
neural-processing-lab.github.io
· LibriBrain Competition 2025
033
Reposted by Julen Etxaniz
Eneko Agirre @eagirre.bsky.social · 15/10/2025
Very interesting dataset related to a yet unsolved problem: how do children learn language. Basque and other 44 languages! Glad that our center is involved with co-author @juletxara.bsky.social
041
Reposted by Julen Etxaniz
Jaap Jumelet @jumelet.bsky.social · 15/10/2025
🌍Introducing BabyBabelLM: A Multilingual Benchmark of Developmentally Plausible Training Data! LLMs learn from vastly more data than humans ever experience. BabyLM challenges this paradigm by focusing on developmentally plausible data We extend this effort to 45 new languages!
14416
Reposted by Julen Etxaniz
Francesca Padovani @frap98.bsky.social · 14/10/2025
𝐃𝐨 𝐲𝐨𝐮 𝐫𝐞𝐚𝐥𝐥𝐲 𝐰𝐚𝐧𝐭 𝐭𝐨 𝐬𝐞𝐞 𝐰𝐡𝐚𝐭 𝐦𝐮𝐥𝐭𝐢𝐥𝐢𝐧𝐠𝐮𝐚𝐥 𝐞𝐟𝐟𝐨𝐫𝐭 𝐥𝐨𝐨𝐤𝐬 𝐥𝐢𝐤𝐞? 🇨🇳🇮🇩🇸🇪 Here’s the proof! 𝐁𝐚𝐛𝐲𝐁𝐚𝐛𝐞𝐥𝐋𝐌 is the first Multilingual Benchmark of Developmentally Plausible Training Data available for 45 languages to the NLP community 🎉 arxiv.org/abs/2510.10159
24016
Reposted by Julen Etxaniz
Mikel Zubillaga @mikel-zubillaga.bsky.social · 26/09/2025
Ayer presenté mi proyecto de tesis "Document Level Information Extraction in Low-Resource Languages" en el simposio doctoral del congreso #SEPLN2025 ¡Eskerrik asko a los organizadores!
043
Reposted by Julen Etxaniz
Imanol Miranda @imirandam.bsky.social · 18/06/2025
#newHitzPaper Can a simple inference-time approach unlock better Vision-Language Compositionality?🤯 Our latest paper shows how adding structure at inference significantly boosts performance in popular dual-encoder VLMs on different datasets. Read more: arxiv.org/abs/2506.09691
163
Reposted by Julen Etxaniz
UEU @ueu-orokorra.bsky.social · 30/05/2025
Jakintza alor guztietako mikroaurkezpenen txanda kafera bitarte #ikergazte2025
043
Reposted by Julen Etxaniz
Aholab Taldea @aholab.bsky.social · 30/05/2025
Primeran aritu dire gure kideak #IKERGAZTE2025 kongresuan!
033
Reposted by Julen Etxaniz
Sagnik Mukherjee @sagnikmukherjee.bsky.social · 21/05/2025
🚨 Paper Alert: “RL Finetunes Small Subnetworks in Large Language Models” From DeepSeek V3 Base to DeepSeek R1 Zero, a whopping 86% of parameters were NOT updated during RL training 😮😮 And this isn’t a one-off. The pattern holds across RL algorithms and models. 🧵A Deep Dive
1113
Reposted by Julen Etxaniz
Joxerra Bengoetxea @bengoetxeaehu.bsky.social · 24/03/2025
Euskal Herriko Unibertsitateak #HITZ bezalako erreferentziazko ikerketa zentroak izateak harrotasunez betetzen gaitu. Eneko Agirre zuzendaria eta bere taldekideekin izan gara goizean. Jarrai dezazuela bide berriak urratzen hizkuntza eta ahotsa ardatz duen AAren inguruan, @hitz-zentroa.bsky.social.
0156
Reposted by Julen Etxaniz
Ahmed S. M. Elhady @ahmedelhady.bsky.social · 26/02/2025
🧙‍♂️ New paper 🧙‍♀️: Presenting Wicked: a simple automated method to make MCQA benchmarks more challenging. Wicked shook up 18 open-weight LLMs on 6 benchmarks, with up to 19.7% performance drop with direct prompting 🤯 Paper: shorturl.at/1CGq0 Code: shorturl.at/n2nCU
186
Reposted by Julen Etxaniz
Blanca Calvo Figueras @blanca-calvo-figue.bsky.social · 17/02/2025
Are humans able to assess if an argument is valid? Can LLMs help them raise the appropriate critical questions to assess its validity? 🤖 Let's use AI to make us more critical!
155
Reposted by Julen Etxaniz
HiTZ zentroa (EHU) @hitz-zentroa.bsky.social · 17/02/2025
Bisita izan dugu! Eusko Jaurlaritzako Hizkuntza Politikarako Sailburuordetzako talde zabalarekin, Aitor Aldasoro sailburuordea, Josune Irabien eta Sonia Rodriguez barne, arratsalde ederra pasa genuen zentroa erakusten eta Ikergaitu proiektuko azken emaitzak erakusten www.hitz.eus/iker-gaitu
045
Reposted by Julen Etxaniz
HiTZ zentroa (EHU) @hitz-zentroa.bsky.social · 17/02/2025
Bisita izan dugu! Zientzia, Unibertsitateak eta Berrikuntza saileko Juan Ignacio Pérez sailburua @juanignacioperez.bsky.social , Adolfo Morais eta Xabier Aizpurua sailburuordeekin eguerdi ederra pasa @hitz-zentroa.bsky.social erakusten eta Adimen Artifizialean azken garapenen inguruan jarduten.
044
Reposted by Julen Etxaniz
HiTZ zentroa (EHU) @hitz-zentroa.bsky.social · 23/01/2025
Iker de la Iglesiak medikuntzako argumentuen ebaluazio automatikoari buruzko posterra aurkeztu du #COLING2025 konferentzian
083
Reposted by Julen Etxaniz
HiTZ zentroa (EHU) @hitz-zentroa.bsky.social · 22/01/2025
We are in #COLING2025!
093
Reposted by Julen Etxaniz
HiTZ zentroa (EHU) @hitz-zentroa.bsky.social · 22/01/2025
Paula Ontalvillak "Improving the Efficiency of Visually Augmented Language Models" lana aurkeztu du COLING konferentzian
154
Reposted by Julen Etxaniz
HiTZ zentroa (EHU) @hitz-zentroa.bsky.social · 26/12/2024
We presented two posters at #NeurIPS2024! @gazkune.bsky.social @juletxara.bsky.social and Imanol Miranda
153
Reposted by Julen Etxaniz
Gorka Azkune @gazkune.bsky.social · 08/12/2024
If you are interested in LLMs' knowledge about local cultures, check this paper by @juletxara.bsky.social and come to discuss with us in Vancouver #NeurIPS2024 arxiv.org/abs/2406.07302
arxiv.org
BertaQA: How Much Do Language Models Know About Local Culture?
Large Language Models (LLMs) exhibit extensive knowledge about the world, but most evaluations have been limited to global or anglocentric subjects. This raises the question of how well these models p...
064
Reposted by Julen Etxaniz
Ai2 @ai2.bsky.social · 26/11/2024
Meet OLMo 2, the best fully open language model to date, including a family of 7B and 13B models trained up to 5T tokens. OLMo 2 outperforms other fully open models and competes with open-weight models like Llama 3.1 8B — As always, we released our data, code, recipes and more 🎁
The OLMo 2 models sit at the Pareto frontier of training FLOPs vs model average performance.
515235
Reposted by Julen Etxaniz
Luca Soldaini 🎀 @soldaini.net · 26/11/2024
OLMo 2 is out 🥳 7B and 13B trained on 5T tokens, and meticulousy instruction tuned using Tulu 3 recipe. Simply the best fully open models yet. Really proud of the work & the amazing team at @ai2.bsky.social
926044