StepFun-Prover Preview: Let's Think and Verify Step by Step
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913987415769088 |
|---|---|
| author | Shang, Shijie Wan, Ruosi Peng, Yue Wu, Yutong Chen, Xiong-hui Yan, Jie Zhang, Xiangyu |
| author_facet | Shang, Shijie Wan, Ruosi Peng, Yue Wu, Yutong Chen, Xiong-hui Yan, Jie Zhang, Xiangyu |
| contents | We present StepFun-Prover Preview, a large language model designed for formal theorem proving through tool-integrated reasoning. Using a reinforcement learning pipeline that incorporates tool-based interactions, StepFun-Prover can achieve strong performance in generating Lean 4 proofs with minimal sampling. Our approach enables the model to emulate human-like problem-solving strategies by iteratively refining proofs based on real-time environment feedback. On the miniF2F-test benchmark, StepFun-Prover achieves a pass@1 success rate of $70.0\%$. Beyond advancing benchmark performance, we introduce an end-to-end training framework for developing tool-integrated reasoning models, offering a promising direction for automated theorem proving and Math AI assistant. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_20199 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | StepFun-Prover Preview: Let's Think and Verify Step by Step Shang, Shijie Wan, Ruosi Peng, Yue Wu, Yutong Chen, Xiong-hui Yan, Jie Zhang, Xiangyu Artificial Intelligence We present StepFun-Prover Preview, a large language model designed for formal theorem proving through tool-integrated reasoning. Using a reinforcement learning pipeline that incorporates tool-based interactions, StepFun-Prover can achieve strong performance in generating Lean 4 proofs with minimal sampling. Our approach enables the model to emulate human-like problem-solving strategies by iteratively refining proofs based on real-time environment feedback. On the miniF2F-test benchmark, StepFun-Prover achieves a pass@1 success rate of $70.0\%$. Beyond advancing benchmark performance, we introduce an end-to-end training framework for developing tool-integrated reasoning models, offering a promising direction for automated theorem proving and Math AI assistant. |
| title | StepFun-Prover Preview: Let's Think and Verify Step by Step |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2507.20199 |