Update README.md

README.md

@@ -101,6 +101,13 @@ Those give us the following result
 | RTE   | Accuracy                     | 73.3 |
 
 
+If you use our models, please cite the following papers.
 
-
-
+```
+@article{wang2022pretraining,
+  title={Pretraining Without Attention},
+  author={Wang, Junxiong and Yan, Jing Nathan and Gu, Albert and Rush, Alexander M},
+  journal={arXiv preprint arXiv:2212.10544},
+  year={2022}
+}
+```