Add model card metadata and description

README.md

@@ -1,13 +1,32 @@
 ---
-license: apache-2.0
 base_model:
 - meta-llama/Meta-Llama-3-8B-Instruct
+license: apache-2.0
+library_name: transformers
+pipeline_tag: text-generation
 ---
+
+# Mix Data or Merge Models? RESM-Llama-3-8B
+
+This repository contains the model weights for the 3H-aligned Large Language Model (LLM) presented in the paper [Mix Data or Merge Models? Balancing the Helpfulness, Honesty, and Harmlessness of Large Language Model via Model Merging](https://huggingface.co/papers/2502.06876).
+
+## Description
+
+Achieving a balanced alignment across Helpfulness, Honesty, and Harmlessness (the 3H dimensions) is critical for responsible AI. This model was developed using **RESM** (**R**eweighting **E**nhanced task **S**ingular **M**erging), a novel model merging method that utilizes outlier weighting and sparsity-aware rank selection strategies. 
+
+RESM is designed to address challenges inherent in 3H-aligned merging, such as preference noise accumulation and layer sparsity adaptation. By working at the parameter level, it provides a conflict-resolution strategy that outperforms traditional data mixture methods in achieving balanced LLM alignment.
+
+- **Base Model:** [Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct)
+- **Method:** RESM (Reweighting Enhanced task Singular Merging)
+- **Optimization Goals:** Helpfulness, Honesty, and Harmlessness (3H)
+
 ## Citation
-```
+
+```bibtex
 @article{yang2025mix,
   title={Mix Data or Merge Models? Balancing the Helpfulness, Honesty, and Harmlessness of Large Language Model via Model Merging},
-  author={Yang, Jinluan and Jin, Dingnan and Tang, Anke and Shen, Li and Zhu, Didi and Chen, Zhengyu and Wang, Daixin and Cui, Qing and Zhang, Zhiqiang and Zhou, Jun and others},
+  author={Yang, Jinluan and Jin, Dingnan and Tang, Anke and Shen, Li and Zhu, Didi + and Chen, Zhengyu and Wang, Daixin and Cui, Qing and Zhang, Zhiqiang and Zhou, Jun and Fei Wu and Kun Kuang},
   journal={arXiv preprint arXiv:2502.06876},
   year={2025}
-}
+}
+```