iiinova — where your digital self keeps evolving

Reposted by Oscar Sainz

HiTZ zentroa (EHU) @hitz-zentroa.bsky.social · 31/10/2025

#Latxa txatbota probatarako erabilgarri jarri dugu! latxa.hitz.eus Jaso ditugun eskaerei erantzunez zuen eskura jarri dugu Latxaren bertsio ahaltsuena, chatGPT-tik gertu dabilena, baina euskara txukunagoa sortuz.

latxa.hitz.eus

Gradio

Click to try out the app!

1910

Reposted by Oscar Sainz

HiTZ zentroa (EHU) @hitz-zentroa.bsky.social · 10/07/2025

Ayer uno de nuestros investigadores, Oscar Sainz (@osainz.bsky.social‬), fue galardonado con el premio a la mejor tesis doctoral en Inteligencia Artificial por la Asociación Española para la Inteligencia Artificial (AEPIA). ¡Enhorabuena! 🥳

052

Reposted by Oscar Sainz

BERRIA @berria.eus · 16/06/2025

«Kaixo, Latxa naiz. Zer jakin nahi duzu gaur?». Euskarazko txatbota sortu du EHUko HiTZ ikerketa zentroak. Oraindik ez dute publikora zabaldu, baina garatzaileek eta enpresek eskuratzeko aukera dute. BERRIAko testuak erabili dituzte Latxa entrenatzeko. t.co/OPVNnBG2xW?utm_...

t.co

«Kaixo, Latxa naiz. Zer jakin nahi duzu gaur?»

Euskarazko txatbota sortu du EHUko HiTZ ikerketa zentroak. Oraindik ez dute jendaurrean zabaldu, baina garatzaileek eta enpresek eskuratzeko aukera dute. BERRIAko testuak erabili dituzte Latxa entr...

023

Oscar Sainz @osainz.bsky.social · 11/06/2025

While the experiments were not complicated, they required the collaboration of amazing co-authors, many compute hours, and of course, the impressive collaboration of the Basque community that was involved in manually assessing the models on an arena style evaluation. Thank you!

011

Oscar Sainz @osainz.bsky.social · 11/06/2025

In this work we face the challenge of developing instruct models for Basque, a low-resource language. Continue pretraining base models is intuitive, but what about instructed models? We analyze systematically all different approaches to find the best solution. 2/3

121

Oscar Sainz @osainz.bsky.social · 11/06/2025

Do you know that you can continue pretraining Instructed LLMs without losing their instruction following capabilities? We did so to teach Basque to Llama models with promising results! Interestingly, you only need English instructions and target language corpora 🤯 1/3

163

Reposted by Oscar Sainz

HiTZ zentroa (EHU) @hitz-zentroa.bsky.social · 11/06/2025

[1/7] #newHitzPaper Many languages are underserved by open LLMs, and face the following question: Which is the best way to produce open instruction-tuned LLMs for low-resource languages? We obtained great results for a cost-effective option! 📄Paper: arxiv.org/abs/2506.07597

173