Link model card to paper and fix code block formatting

This PR improves the model card by adding an explicit link to the paper [The Illusion of High Utility in Safety Alignment of Text-to-Image Diffusion Models](https://huggingface.co/papers/2607.00402). Linking the paper will automatically associate this repository with its corresponding Hugging Face paper page. It also ensures that the Python code block in the usage section is properly closed.

README.md

@@ -1,17 +1,17 @@
 ---
-license: mit
 library_name: diffusers
+license: mit
+pipeline_tag: text-to-image
 tags:
 - text-to-image
 - safety-alignment
 - stable-diffusion
 - ECCV
-pipeline_tag: text-to-image
 ---
 
 # SAGE: Structure-Aware Geometric Regularization (ECCV-26)
 
-**Paper:** The Illusion of High Utility in Safety Alignment of Text-to-Image Diffusion Models  
+**Paper:** [The Illusion of High Utility in Safety Alignment of Text-to-Image Diffusion Models](https://huggingface.co/papers/2607.00402)  
 **Authors:** Adeel Yousaf, Soumik Ghosh, James Beetham, Amrit Singh Bedi, Mubarak Shah  
 **Institution:** University of Central Florida  
 **Project Page:** [https://adeelyousaf.github.io/SAGE_ECCV26_Project_Page/](https://adeelyousaf.github.io/SAGE_ECCV26_Project_Page/)
@@ -42,3 +42,4 @@ pipe.text_encoder.load_state_dict(torch.load(ckpt_path, map_location="cpu"))
 
 pipe = pipe.to("cuda")
 image = pipe("a photo of a dog in a park").images[0]
+```