analiza aiqbench
This paper presents open-weight language models (0.6B to 14B parameters) adapted to Wolof, a language with about 23M speakers that current open models support poorly. The authors use instruction tuning, preference alignment, and model merging, and report results on 63 tasks, with the 8B and 14B models competitive with much larger commercial systems. For AI readers, the main takeaways are the reusable adaptation recipe for low-resource languages and the public release, though the benchmark claims and the status of the customer-care pilots still need independent verification.
Wolof-Adapted Industry-Ready Open-Weight Language Models
arXiv:2610.20824v1 Announce Type: cross Abstract: Wolof is spoken by 23M people, mostly in Senegal, but remains weakly supported by open-weight language models, while API models raise cost, governance, and sovereignty constraints for local deployments. In this paper, we present 0.6B-14B open-weight models adapted to Wolof through instruction tuning, preference alignment, and model merging. After presenting the pipeline and data, extensive experiments on 63~public and in-house tasks show good performance on Wolof understanding and generation while remaining effective on other languages and cross-lingual skills, with the 8B and 14B models being particularly competitive with substantially larger commercial systems. The models are currently being evaluated in internal customer-care pilots and are publicly available at https://huggingface.co/collections/Orange/wolof-lm .
Abstract:Wolof is spoken by 23M people, mostly in Senegal, but remains weakly supported by open-weight language models, while API models raise cost, governance, and sovereignty constraints for local deployments. In this paper, we present 0.6B-14B open-weight models adapted to Wolof through instruction tuning, preference alignment, and model merging. After presenting the pipeline and data, extensive experiments on 63~public and in-house tasks show good performance on Wolof understanding and generation while remaining effective on other languages and cross-lingual skills, with the 8B and 14B models being particularly competitive with substantially larger commercial systems. The models are currently being evaluated in internal customer-care pilots and are publicly available at this https URL .
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
arXiv:2610.20824 [cs.CL]
(or arXiv:2610.20824v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2610.20824 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 209 cuvinte
Subiecte conexe
NVIDIA GTC Taipei at COMPUTEX: Live Updates on What’s Next in AI
NVIDIA · 21 mai 2026
Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Latent Space · 13 oct. 2026
NVIDIA Nemotron 3 Nano Omni Powers Multimodal Agent Reasoning in a Single Efficient Open Model
NVIDIA · 28 apr. 2026
Why Telecom Operators Are Building Their AI Strategy on Open Models
NVIDIA · 6 oct. 2026