hkust-nlp/Qwen-2.5-Math-7B-SimpleRL

Simple Reinforcement Learning for Reasoning

This is the model checkpoint in Project SimpleRL. Qwen-2.5-Math-7B-SimpleRL is the simple RL training from the base model with initial warmup stage.

Quick Start

Please generate content using the following template:

"o1_cot": (
    '[Round 0] USER:\\n{input}\\nPlease reason step by step, and put your final answer within \\\\boxed{{}}. ASSISTANT:\\n',
    "{output}",
    "\\n\\n"
)

Alternatively, you can use our evaluation code and specify the Prompt type as "o1_cot"

Citation

If you find this blog or our code useful, we would appreciate it if you could cite our work:

@misc{
    zeng2025simplerl,
    title={7B Model and 8K Examples: Emerging Reasoning with Reinforcement Learning is Both Effective and Efficient},
    author={Weihao Zeng and Yuzhen Huang and Wei Liu and Keqing He and Qian Liu and Zejun Ma and Junxian He},
    year={2025},
    howpublished={\url{https://hkust-nlp.notion.site/simplerl-reason}},
    note={Notion Blog}
}