NbAiLab/borealis-open-12b
NbAiLab/borealis-open-12b is a 12-billion parameter instruction-tuned language model from the National Library of Norway (NbAiLab), based on Google's Gemma 3 architecture. This model is specifically fine-tuned for Norwegian-centric instruction following, excelling in tasks like drafting, summarization, and question answering in Norwegian. It is an "open release" variant, meaning it does not include certain copyright-protected newspaper material used in the full Borealis models, making it suitable for general Norwegian language applications.
Loading preview...
Borealis Open 12B: Norwegian-Centric Instruction-Tuned Model
NbAiLab/borealis-open-12b is a 12-billion parameter instruction-tuned model developed by the National Library of Norway (NbAiLab). It is part of the Borealis family and is an "open release" variant, meaning it excludes certain copyright-protected data used in the full Borealis models. This model is built upon the Gemma 3 architecture from Google (google/gemma-3-12b-it) and has been specifically fine-tuned for Norwegian-centric instruction following.
Key Capabilities
- Norwegian Language Proficiency: Optimized for assistant-style tasks, writing, summarization, and question answering in Norwegian, including Bokmål and Nynorsk.
- Instruction Following: Fine-tuned with a supervised fine-tuning (SFT) dataset, NbAiLab/aurora-sft-open, prepared by the National Library of Norway for Norwegian-specific behaviors.
- Safety Alignment: Incorporates "prompt baking" for safety alignment, balancing model quality, usefulness, and safer behavior, though outputs can still be incorrect or biased.
Good For
- Norwegian Assistant Applications: Ideal for developing chatbots or virtual assistants that require strong Norwegian language understanding and generation.
- Content Generation & Summarization: Useful for drafting and summarizing texts, and generating news article titles and ingress texts in Norwegian.
- Language Assessment: Suitable for assessing and improving Norwegian writing style and quality, and evaluating language coverage for Norwegian dialects.
- Research & Development: Provides a robust base for further research into Norwegian language models, particularly for those requiring an open-source model without copyright-restricted data.