Add CRJudgeBench paper citation to dataset card

README.md

@@ -122,3 +122,19 @@ Treat `true` as the positive class. Report accuracy, macro-F1, per-class precisi
 No dataset-wide license is specified for this release. This card does not assign a new license to the collected code or discussions. Source repositories are identified in the `repo` field, and PR numbers are available in `pull_request.pull_number`.
 
 When referencing the dataset, link to [dcloud347/CRJudgeBenchmark](https://huggingface.co/datasets/dcloud347/CRJudgeBenchmark) and record the dataset revision used for the experiment.
+
+## Citation
+
+If you use CRJudgeBenchmark in your research, please cite our paper:
+
+```bibtex
+@misc{pan2026crjudgebenchaidetectplausible,
+      title={CRJudgeBench: Can AI Detect Plausible but Invalid Code Reviews?},
+      author={Yue Pan and Jiawei Li and Ziyuan Zhang and Xiangxin Zhao and He Ye},
+      year={2026},
+      eprint={2609.37216},
+      archivePrefix={arXiv},
+      primaryClass={cs.SE},
+      url={https://arxiv.org/abs/2609.37216},
+}
+```