| <h1 align="center">To Mix or To Merge?<br/>Toward Multi-Domain Reinforcement Learning for Large Language Models</h1> |
|
|
| <p align="center"> |
| <a href="https://arxiv.org/abs/2602.12566"><img src="https://img.shields.io/badge/arXiv-Paper-b31b1b.svg" alt="arXiv"></a> |
| <a href="https://hf.co/collections/Jackwang111/m2rl"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Models-ffd21e.svg" alt="Hugging Face"></a> |
| <a href="https://modelscope.cn/collections/whq1111/M2RL"><img src="https://img.shields.io/badge/ModelScope-Models-5f4bff.svg" alt="ModelScope"></a> |
| <a href="https://github.com/Mosi-AI/M2RL"><img src="https://img.shields.io/badge/GitHub-Code-181717.svg?logo=github" alt="GitHub"></a> |
| <a href="https://openreview.net/forum?id=jP7j5XkG8J"><img src="https://img.shields.io/badge/COLM-2026-4b8bbe.svg" alt="COLM 2026"></a> |
| </p> |
|
|
| <p align="center"> |
| <b>Haoqing Wang<sup>†</sup>, Xiang Long<sup>†</sup>, Ziheng Li<sup>†</sup>, Yilong Xu, Tingguang Li, Yehui Tang<sup>✉</sup></b><br/> |
| Samsung Research, Beijing, China · Peking University |
| </p> |
|
|
| --- |
|
|
| ## 📰 News |
|
|
| - **[2026.09.07]** 🎉 The model checkpoints are now open-sourced on [Hugging Face](https://hf.co/collections/Jackwang111/m2rl) and [ModelScope](https://modelscope.cn/collections/whq1111/M2RL)! **Feel free to use our checkpoints for your post-training research (e.g., weight merging, multi-teacher on-policy distillation)!** |
| - **[2026.07.09]** 🎉 Our paper is accepted to **COLM 2026**! |
|
|
| --- |
|
|
| ## 📚 Citation |
|
|
| If you find this work useful, please consider citing: |
|
|
| ```bibtex |
| @inproceedings{ |
| wang2026to, |
| title={To Mix or To Merge: Toward Multi-Domain Reinforcement Learning for Large Language Models}, |
| author={Haoqing Wang and Xiang Long and Ziheng Li and Yilong Xu and Tingguang Li and Yehui Tang}, |
| booktitle={Third Conference on Language Modeling}, |
| year={2026}, |
| url={https://openreview.net/forum?id=jP7j5XkG8J} |
| } |
| ``` |
|
|
