NbAiLab/borealis-open-1b
NbAiLab/borealis-open-1b is a 1-billion parameter instruction-tuned causal language model developed by the National Library of Norway (NbAiLab), based on the Gemma 3 family. 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, meaning it does not include data from specific rights-holder agreements, and is intended for generalist Norwegian language applications.
Loading preview...
Borealis Open 1B: Norwegian-Centric Instruction-Tuned Model
NbAiLab/borealis-open-1b is a 1-billion parameter instruction-tuned model from the National Library of Norway (NbAiLab), built upon the Gemma 3 architecture. This model is part of the Borealis family, specifically designed for Norwegian language tasks.
Key Capabilities
- Norwegian-centric instruction following: Fine-tuned for assistant-style tasks, writing, summarization, and question answering in Norwegian.
- Open Release: This version (
-opensuffix) does not incorporate data from specific rights-holder agreements, making it broadly accessible. - Safety Alignment: Utilizes prompt baking and weighted merging of SFT and aligned models to balance quality, usefulness, and safer behavior, though outputs can still be incorrect or biased.
- Multilingual Evaluation: Evaluated with NorEval, MMLU-English, and nb-gpt-bench, demonstrating performance across relevant benchmarks.
Good For
- Norwegian language applications: Ideal for drafting, summarizing, and answering questions in Norwegian, including Bokmål and Nynorsk.
- Assessment of Norwegian writing: Useful for evaluating and improving Norwegian writing style and quality.
- Research and development: Provides a foundation for exploring behavior and language coverage for Norwegian dialects.
- Generalist use cases: Intended for a wide range of applications where Norwegian language understanding and generation are critical.