Add pipeline tag and update paper reference

This PR improves the model card by:
- Adding the `pipeline_tag: text-to-image` to the metadata, ensuring the model is discoverable under the text-to-image pipeline on the Hugging Face Hub.
- Updating the main paper reference in the model card to [AlignGuard: Scalable Safety Alignment for Text-to-Image Generation](https://huggingface.co/papers/2412.10493), which is the paper explicitly linked in the prompt.

README.md

@@ -1,12 +1,14 @@
 ---
 library_name: diffusers
 license: mit
+pipeline_tag: text-to-image
 ---
-This is the official released checkpoint of **an aligned SDXL v1.0** of SafetyDPO: Scalable Safety Alignment for Text-to-Image Generation, designed to generate more safe images from our Safe-SDXL. 
 
-Our project page is [🏠SafetyDPO HomePage](https://safetydpo.github.io/) and the GitHub repo is [⚙️SafetyDPO GitHub](https://github.com/Visualignment/SafetyDPO) where we released all the code and the data. 
+This is the official released checkpoint of **an aligned SDXL v1.0** from the paper [AlignGuard: Scalable Safety Alignment for Text-to-Image Generation](https://huggingface.co/papers/2412.10493), designed to generate more safe images from our Safe-SDXL.
 
-In the future, we will release additional safe models. 
+Our project page is [🏠SafetyDPO HomePage](https://safetydpo.github.io/) and the GitHub repo is [⚙️SafetyDPO GitHub](https://github.com/Visualignment/SafetyDPO) where we released all the code and the data.
+
+In the future, we will release additional safe models.
 
 # Usage
 A simple use case of our model is: