Text Classification
PEFT
Safetensors
English
safety
guardrail
constitutional-classifiers
jailbreak-detection
content-moderation
Instructions to use EuroSafeAI/EuroSafeGuard-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use EuroSafeAI/EuroSafeGuard-v1 with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
Card: Qwen3.5-9B license; note on the five runs
Browse files
README.md
CHANGED
|
@@ -35,6 +35,8 @@ categories.
|
|
| 35 |
|
| 36 |
## Quickstart
|
| 37 |
|
|
|
|
|
|
|
| 38 |
```python
|
| 39 |
from eurosafeguard import EuroSafeGuard # pip install "eurosafeguard @ git+https://github.com/EuroSafeAI/EuroSafeGuard"
|
| 40 |
# or download eurosafeguard.py from this repo
|
|
@@ -101,7 +103,7 @@ End to end (deployed cascade in front of Llama-3-8B-Instruct, 1,974 held-out Jai
|
|
| 101 |
|
| 102 |
Apache-2.0 for our adapters, heads, probe and code. Base models keep their licenses:
|
| 103 |
[Llama 3.1 Community License](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) and
|
| 104 |
-
[
|
| 105 |
|
| 106 |
## Citation
|
| 107 |
|
|
|
|
| 35 |
|
| 36 |
## Quickstart
|
| 37 |
|
| 38 |
+
The repo has five input-classifier training runs (`seed=0` to `4`). The paper reports their mean. Runs differ on search attacks (11–48% of PAIR behaviors broken, run 0 is the weakest), so choose one with `seed=k`.
|
| 39 |
+
|
| 40 |
```python
|
| 41 |
from eurosafeguard import EuroSafeGuard # pip install "eurosafeguard @ git+https://github.com/EuroSafeAI/EuroSafeGuard"
|
| 42 |
# or download eurosafeguard.py from this repo
|
|
|
|
| 103 |
|
| 104 |
Apache-2.0 for our adapters, heads, probe and code. Base models keep their licenses:
|
| 105 |
[Llama 3.1 Community License](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) and
|
| 106 |
+
[Apache-2.0 for Qwen3.5-9B](https://huggingface.co/Qwen/Qwen3.5-9B).
|
| 107 |
|
| 108 |
## Citation
|
| 109 |
|