kashef-marbert-base-augmented

This model is part of the Kashef research project for detecting machine-generated Arabic text.

Model Details

• Base Architecture: UBC-NLP/MARBERT
• Dataset: Augmented Dataset (Original + External)
• Description: Trained on a large combined dataset including external sources.

Performance (Blind Test)

Metric	Score
F1-Score	0.636
Accuracy	57%