Update README.md

README.md

@@ -108,5 +108,13 @@ NoisyCLIP is trained with the standard CLIP contrastive objective on pairs of **
 If you find NoisyCLIP useful for your research, please cite:
 
 ```bibtex
-TO BE UPDATED
+@misc{ramos2026earlyestimationlanguagelatent,
+      title={Early Estimation of Language to Latent Alignment in Diffusion Models}, 
+      author={Vasco Ramos and Regev Cohen and Idan Szpektor and Joao Magalhaes},
+      year={2026},
+      eprint={2512.08505},
+      archivePrefix={arXiv},
+      primaryClass={cs.CV},
+      url={https://arxiv.org/abs/2512.08505}, 
+}
 ```