Add paper link, project links, and update metadata

README.md

@@ -1,22 +1,28 @@
 ---
-license: apache-2.0
+base_model:
+- Qwen/Qwen2.5-VL-7B-Instruct
 language:
 - en
+license: apache-2.0
 metrics:
 - accuracy
-base_model:
-- Qwen/Qwen2.5-VL-7B-Instruct
+library_name: transformers
+pipeline_tag: image-text-to-text
 ---
+
 <p align="center">
   <a href="https://huggingface.co/Carol0110/UniRM/blob/main/README.md"><b>English</b></a> | <a href="https://huggingface.co/Carol0110/UniRM/blob/main/README_zh.md">中文</a>
 </p>
 
 # UniRM: Multi-Head Scalar Reward Model for Multimodal Moderation
 
-**UniRM** is a **multi-head scalar reward model** that provides **interpretable, attribute-level scoring** for multimodal moderation.  
-It is designed to support policy optimization for open-ended reasoning in **UniMod**, especially for the posterior response stage where deterministic labels are absent.
+**UniRM** is a **multi-head scalar reward model** that provides **interpretable, attribute-level scoring** for multimodal moderation. It was introduced in the paper [From Sparse Decisions to Dense Reasoning: A Multi-attribute Trajectory Paradigm for Multimodal Moderation](https://huggingface.co/papers/2602.02536).
+
+- **Project Page:** [trustworthylab.github.io/UniMod/](https://trustworthylab.github.io/UniMod/)
+- **Repository:** [github.com/Carol-gutianle/UniMod](https://github.com/Carol-gutianle/UniMod)
+- **Paper:** [arXiv:2602.02536](https://huggingface.co/papers/2602.02536)
 
-UniRM decouples reward attribution into multiple dimensions so the model can distinguish **stylistic quality** from **safety boundaries** (privacy, bias, toxicity, legality), enabling transparent diagnosis and stable optimization.
+UniRM is designed to support policy optimization for open-ended reasoning in **UniMod**, especially for the posterior response stage where deterministic labels are absent. It decouples reward attribution into multiple dimensions so the model can distinguish **stylistic quality** from **safety boundaries** (privacy, bias, toxicity, legality), enabling transparent diagnosis and stable optimization.
 
 ---
 
@@ -25,7 +31,7 @@ UniRM decouples reward attribution into multiple dimensions so the model can dis
 > UniRM demo video:
 
 <video controls preload="metadata" style="width:100%; max-width:900px; border-radius:12px;">
-  <source src="https://huggingface.co/Carol0110/UniRM/blob/main/unirm.mp4" type="video/mp4">
+  <source src="https://huggingface.co/Carol0110/UniRM/resolve/main/unirm.mp4" type="video/mp4">
 </video>
 
 ---
@@ -39,4 +45,19 @@ git clone https://github.com/TideDra/lmm-r1.git
 cd lmm-r1
 pip install -e .[vllm]
 pip install flash_attn --no-build-isolation
-python unirm.py --model_path {PATH_TO_UNIRM}
+python unirm.py --model_path {PATH_TO_UNIRM}
+```
+
+## Citation
+
+```bibtex
+@misc{gu2026sparsedecisionsdensereasoning,
+      title={From Sparse Decisions to Dense Reasoning: A Multi-attribute Trajectory Paradigm for Multimodal Moderation}, 
+      author={Tianle Gu and Kexin Huang and Lingyu Li and Ruilin Luo and Shiyang Huang and Zongqi Wang and Yujiu Yang and Yan Teng and Yingchun Wang},
+      year={2026},
+      eprint={2602.02536},
+      archivePrefix={arXiv},
+      primaryClass={cs.LG},
+      url={https://arxiv.org/abs/2602.02536}, 
+}
+```