Model Details

Best Mistral-7B-Instruct-v0.3 checkpoint unlearned using RR with the Keyword-Cyber forget set. For more details, please check our paper.

sources

Performance

WMDP-Cyber tinyMMLU GSM8k TriviaQA
Mistral-7B-Instruct-v0.3 41.52 64.20 50.19 56.81
Mistral-7B-Instruct-v0.3_RR_Keyword-Cyber 23.60 63.44 45.71 56.77

Citation

If you find this useful in your research, please consider citing our paper:

@misc{zhu2025llmunlearningexpertcurated,
      title={LLM Unlearning Without an Expert Curated Dataset}, 
      author={Xiaoyuan Zhu and Muru Zhang and Ollie Liu and Robin Jia and Willie Neiswanger},
      year={2025},
      eprint={2508.06595},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2508.06595}, 
}
Downloads last month
5
Safetensors
Model size
10B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Collection including WhyTheMoon/Mistral-7B-Instruct-v0.3_RR_Keyword-Cyber