Transformers
English
Hindi
Panjabi
safety
activation-steering
refusal
interpretability
code-mixing
romanized
aya
Instructions to use sunit12/tiny-aya-global-codemix-refusal-guard with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use sunit12/tiny-aya-global-codemix-refusal-guard with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("sunit12/tiny-aya-global-codemix-refusal-guard", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Commit History
Yesterday
-
4-bit + Fire results, optional gate_v3_1
Verified signature8125763 -
Gate v3 (layer 13): broader categories, romanized everyday negatives; usable-refusal results
Verified signature8db69b0 -
Model card with held-out results
Verified signatureb51808c -
Guard: gate_v2 (L9) + refusal vectors + guard.py
Verified signature90588ec -
initial commit
Verified signature192992a