File size: 895 Bytes
fa405a0 99498f5 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 |
---
{}
---
# HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models
This model is a Guard Model, specifically designed to classify the safety of LLM conversations.
It is fine-tuned from DeBERTa-v3-large and trained using **HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models**.
The training process involves knowledge distillation paired with data augmentation, using our [**HarmAug Generated Dataset**](https://huggingface.co/datasets/AnonHB/HarmAug_generated_dataset).
For more information, please refer to our [anonymous github](https://anonymous.4open.science/r/HarmAug/)
|