Official implementation of paper:

"Are LLMs Reliable Rankers? Rank Manipulation via Two-Stage Token Optimization" Tiancheng Xing, Jerry Li, Yixuan Du, Xiyang Hu

We've prepared scripts to quickly run the code. Simply run the script like:

bash run_raf.sh

The evaluation script helps to evaluate adversarial prompts generated by different approaches:

bash evaluate.sh

If you find our code useful for your research, please cite our paper.

@article{xing2025llms,
  title={Are LLMs Reliable Rankers? Rank Manipulation via Two-Stage Token Optimization},
  author={Xing, Tiancheng and Li, Jerry and Du, Yixuan and Hu, Xiyang},
  journal={arXiv preprint arXiv:2510.06732},
  year={2025}
}