38 lines
1.5 KiB
Markdown
38 lines
1.5 KiB
Markdown
|
---
|
|||
|
|
license: apache-2.0
|
|||
|
|
base_model:
|
|||
|
|
- Qwen/Qwen2.5-1.5B-Instruct
|
|||
|
|
tags:
|
|||
|
|
- safety
|
|||
|
|
- alignment
|
|||
|
|
---
|
|||
|
|
|
|||
|
|
# Qwen2.5-1.5B-Instruct-EASE
|
|||
|
|
This model is a fine-tuned version of [Qwen/Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct) on the [EASE-SafetyReasoning](https://huggingface.co/datasets/HaonanShi/EASE-STAR41K-SafetyReasoning-10K) dataset.
|
|||
|
|
|
|||
|
|
## Model description
|
|||
|
|
This is the safety reasoning aligned version model under the framework,[**EASE**](https://arxiv.org/pdf/2511.06512). We fine-tune Qwen2.5-1.5B-Instruct to enable **adaptive safety reasoning activation**. The model triggers explicit safety reasoning only under jailbreak-like semantics, while avoiding unnecessary safety reasoning on benign or general prompts. This design aims to maintain the model’s general task effectiveness and efficiency, while improving robustness against jailbreak attacks.
|
|||
|
|
|
|||
|
|
## Intended use
|
|||
|
|
Safety-oriented research on:(1)Safety alignment, (2)Small language models and (3)Jailbreak robustness
|
|||
|
|
|
|||
|
|
## Citation
|
|||
|
|
If our model could help you, please cite our paper, thanks!🤗
|
|||
|
|
|
|||
|
|
**EASE: Practical and Efficient Safety Alignment for Small Language Models(AAAI26)(oral)**
|
|||
|
|
|
|||
|
|
https://arxiv.org/pdf/2511.06512
|
|||
|
|
|
|||
|
|
```bibtex
|
|||
|
|
@inproceedings{shi2026ease,
|
|||
|
|
title={Ease: Practical and efficient safety alignment for small language models},
|
|||
|
|
author={Shi, Haonan and Wang, Guoli and Ouyang, Tu and Wang, An},
|
|||
|
|
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
|
|||
|
|
volume={40},
|
|||
|
|
number={44},
|
|||
|
|
pages={37923--37931},
|
|||
|
|
year={2026}
|
|||
|
|
}
|
|||
|
|
```
|
|||
|
|
|