Add library_name, pipeline_tag and link to paper

README.md

@@ -1,14 +1,31 @@
 ---
-license: apache-2.0
 base_model:
 - mistralai/Mistral-7B-Instruct-v0.2
+license: apache-2.0
+library_name: transformers
+pipeline_tag: text-generation
 ---
 
+# RESM-Mistral-7B
+
+This repository contains the model checkpoint for **RESM-Mistral-7B**, as introduced 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
+RESM (**R**eweighting **E**nhanced task **S**ingular **M**erging) is a novel model merging method designed to achieve balanced alignment across the "3H" dimensions: Helpfulness, Honesty, and Harmlessness. 
+
+Traditional data mixture strategies often struggle with conflicting optimization signals. RESM addresses these challenges at the parameter level through:
+1. **Outlier Weighting**: Mitigating preference noise accumulation.
+2. **Sparsity-aware Rank Selection**: Adapting to layer sparsity inherent in aligned LLMs.
+
+This model is a 3H-aligned version based on `mistralai/Mistral-7B-Instruct-v0.2`.
+
 ## Citation
-```
+If you find this work useful, please cite:
+```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},
   journal={arXiv preprint arXiv:2502.06876},
   year={2025}
-}
+}
+```