Add model card and metadata

README.md

@@ -0,0 +1,33 @@
+---
+library_name: transformers
+pipeline_tag: text-classification
+---
+
+# Unified PRM (MAHALO)
+
+This repository contains the Unified Process Reward Model (PRM) introduced in the paper [Simultaneous Multi-objective Alignment Across Verifiable and Non-verifiable Rewards](https://huggingface.co/papers/2510.01167).
+
+## Model Description
+This model is a generalist reward model trained across seven dimensions spanning multiple domains:
+- **Mathematics**: Accuracy and engagement.
+- **Human Values**: Helpfulness, honesty, and truthfulness.
+- **Socratic Mind (AI Tutoring)**: Accuracy and engagement.
+
+It is based on a Llama-3 8B backbone and fine-tuned for sequence classification to provide step-level supervision as part of the **MAHALO** (Multi-Action-Head ALignment with PRM-guided DecOding) framework.
+
+- **Paper:** [Simultaneous Multi-objective Alignment Across Verifiable and Non-verifiable Rewards](https://arxiv.org/abs/2510.01167)
+- **Repository:** [GitHub - pearls-lab/multiobj-align](https://github.com/pearls-lab/multiobj-align)
+
+## Citation
+
+If you find this model useful, please consider citing the original work:
+
+```bibtex
+@article{shen2025simultaneous,
+  title={Simultaneous Multi-objective Alignment Across Verifiable and Non-verifiable Rewards},
+  author={Shen, Yiran and Xia, Yu and Chang, Jonathan and Ammanabrolu, Prithviraj},
+  journal={arXiv preprint arXiv:2510.01167},
+  year={2025},
+  url={https://arxiv.org/abs/2510.01167}
+}
+```