---
license: mit
datasets:
- muse-bench/MUSE-Books
language:
- en
base_model:
- muse-bench/MUSE-books_target
pipeline_tag: text-generation
library_name: transformers
tags:
- unlearn
- machine-unlearning
- llm-unlearning
- data-privacy
- large-language-models
- trustworthy-ai
- trustworthy-machine-learning
- language-model
---

# SimNPO-Unlearned Model on Task "MUSE - Books"

## Model Details

- **Unlearning**:
  - **Task**: [🤗datasets/muse-bench/MUSE-Books](https://huggingface.co/datasets/muse-bench/MUSE-Books)
  - **Method**: [SimNPO](https://arxiv.org/abs/2410.07163)
- **Origin Model**: [🤗muse-bench/MUSE-books_target](https://huggingface.co/muse-bench/MUSE-books_target)
- **Code Base**: [github.com/OPTML-Group/Unlearn-Simple](https://github.com/OPTML-Group/Unlearn-Simple)
- **Research Paper**: ["Simplicity Prevails: Rethinking Negative Preference Optimization for LLM Unlearning"](https://arxiv.org/abs/2410.07163)

## Unlearning Algorithm

This model uses the `SimNPO` unlearning algorithm with the following optimization objective:
$$\ell_{SimNPO}(\mathbf{\theta}) = \mathbb{E}_{(x, y) \in \mathcal{D}_f}\left[-\frac{2}{\beta}\log\sigma\left(-\frac{\beta}{|y|}\log\pi_{\mathbf{\theta}}(y|x) - \gamma\right)\right] + \lambda \mathbb{E}_{(x, y) \in \mathcal{D}_r}[-\log\pi_{\mathbf{\theta}} (y|x)]$$
Unlearning hyper-parameters:
- Learning Rate: `1e-5`
- beta: `0.7`
- lambda: `1.0`
- gamma: `0.0`

## Loading the Model

```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("OPTML-Group/SimNPO-MUSE-Books-iclm-7b", torch_dtype=torch.bfloat16, device_map='auto')
```

## Evaluation Results
||VerbMem Df|KnowMem Df|PrivLeak|KnowMem Dr|
|---|---|---|---|---|
|Origin|99.56|58.32|-56.32|67.01|
|Retrain|14.30|28.90|0.00|74.50|
|NPO|0.00|0.00|-31.17|23.71|
|**SimNPO**|0.00|0.00|-19.82|48.27|

## Citation

If you use this model in your research, please cite: 
```
@article{fan2024simplicity,
  title={Simplicity Prevails: Rethinking Negative Preference Optimization for LLM Unlearning},
  author={Fan, Chongyu and Liu, Jiancheng and Lin, Licong and Jia, Jinghan and Zhang, Ruiqi and Mei, Song and Liu, Sijia},
  journal={arXiv preprint arXiv:2410.07163},
  year={2024}
}
```

## Reporting Issues

Reporting issues with the model: [github.com/OPTML-Group/Unlearn-Simple](https://github.com/OPTML-Group/Unlearn-Simple)