Update model card: add library_name, update pipeline_tag and link paper

README.md

@@ -1,21 +1,25 @@
 ---
-license: apache-2.0
-datasets:
-- agentica-org/DeepScaleR-Preview-Dataset
 base_model:
 - Qwen/Qwen2.5-32B
-pipeline_tag: reinforcement-learning
+datasets:
+- agentica-org/DeepScaleR-Preview-Dataset
+license: apache-2.0
+pipeline_tag: text-generation
+library_name: transformers
 ---
+
 ### Description
-This repository contains the model for [Meta-Awareness Enhances Reasoning Models: Self-Alignment Reinforcement Learning](https://huggingface.co/papers/2510.03259).
+This repository contains the model presented in [Verifying Meta-Awareness via Predictive Rewards in Reasoning Models](https://huggingface.co/papers/2510.03259).
+
+MAPR (Meta-Awareness via Predictive Reward) utilizes a self-generated task of predicting rollout statistics—specifically length, pass-rate, and concepts used—allowing for verification against the actual statistics. By leveraging this self-predictive capability, the model can regulate its reasoning behavior, leading to significant performance gains in accuracy and training efficiency.
 
 ### Official Implementation
-https://github.com/akatigre/MASA-RL
+The official code is available at [https://github.com/akatigre/MASA-RL](https://github.com/akatigre/MASA-RL).
 
 ### Citation
 ```bibtex
 @article{kim2025meta,
-  title={Meta-Awareness Enhances Reasoning Models: Self-Alignment Reinforcement Learning},
+  title={Verifying Meta-Awareness via Predictive Rewards in Reasoning Models},
   author={Kim, Yoonjeon and Jang, Doohyuk and Yang, Eunho},
   journal={arXiv preprint arXiv:2510.03259},
   year={2025}