Update README.md
Browse files
README.md
CHANGED
|
@@ -13,4 +13,44 @@ library_name: transformers
|
|
| 13 |
tags:
|
| 14 |
- conversational
|
| 15 |
- instruct
|
| 16 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
tags:
|
| 14 |
- conversational
|
| 15 |
- instruct
|
| 16 |
+
- experimental
|
| 17 |
+
---
|
| 18 |
+
|
| 19 |
+
# Borealis 4B Instruct (Preview)
|
| 20 |
+
|
| 21 |
+
## Model summary
|
| 22 |
+
**NbAiLab/borealis-4b-instruct-preview** is a **4B-parameter** instruction-tuned **preview** model intended for early testing and feedback. It is an **experiment** and should be treated as pre-release quality.
|
| 23 |
+
|
| 24 |
+
This model is based on [**google/gemma-3-4b-it**](https://huggingface.co/google/gemma-3-4b-it), and fine-tuned on textual instructions only.
|
| 25 |
+
|
| 26 |
+
## Training data
|
| 27 |
+
Supervised fine-tuning (SFT) uses **NbAiLab/aurora-sft-2512** (not released yet).
|
| 28 |
+
|
| 29 |
+
## ⚠️ Safety / alignment disclaimer (important)
|
| 30 |
+
This is a **preview experiment** and **has not been safety-aligned yet**. The model may produce **harmful, biased, or insensitive** outputs (including content that is offensive, unsafe, or inappropriate). Do not use it for safety-critical or high-stakes applications, and add your own safety mitigations if deploying.
|
| 31 |
+
|
| 32 |
+
## Intended use
|
| 33 |
+
- Norwegian-centric assistant-style tasks (e.g., drafting, summarization, Q&A, light reasoning).
|
| 34 |
+
- Assesstment of Norwegian writing style and quality.
|
| 35 |
+
- Early evaluation of behavior, language coverage (Norwegian / Bokmål / Nynorsk), and quality.
|
| 36 |
+
|
| 37 |
+
## Limitations
|
| 38 |
+
- Preview quality; outputs may be unstable and may hallucinate.
|
| 39 |
+
- Not aligned for safety; may follow harmful instructions or generate problematic content (see disclaimer above).
|
| 40 |
+
|
| 41 |
+
## Weights & formats
|
| 42 |
+
|
| 43 |
+
### Transformers (original)
|
| 44 |
+
- **NbAiLab/borealis-4b-instruct-preview** (safetensors).
|
| 45 |
+
|
| 46 |
+
### GGUF quantizations
|
| 47 |
+
Available in [**NbAiLab/borealis-4b-instruct-preview-gguf**](https://huggingface.co/NbAiLab/borealis-4b-instruct-preview-gguf):
|
| 48 |
+
- `model-q8_0.gguf`
|
| 49 |
+
- `model-f16.gguf`
|
| 50 |
+
- `model-bf16.gguf`
|
| 51 |
+
|
| 52 |
+
### MLX (Apple Silicon)
|
| 53 |
+
Available in [**NbAiLab/borealis-4b-instruct-preview-mlx**](https://huggingface.co/NbAiLab/borealis-4b-instruct-preview-mlx)
|
| 54 |
+
|
| 55 |
+
## Acknowledgements
|
| 56 |
+
Thanks to the **Gemma** team at Google for releasing Gemma 3 and to everyone contributing feedback on this preview.
|