StepFun-Formalizer: Unlocking the Autoformalization Potential of LLMs through Knowledge-Reasoning Fusion
Yutong Wu, Di Huang, Ruosi Wan, Yue Peng, Shijie Shang, Chenrui Cao, Lei Qi, Rui Zhang, Zidong Du, Jie Yan, Xing Hu · 2025
Read StepFun-Formalizer: Unlocking the Autoformalization Potential of LLMs through Knowledge-Reasoning Fusion by Yutong Wu, Di Huang, Ruosi Wan, Yue Peng, Shijie Shang, Chenrui Cao, Lei Qi, Rui Zhang, Zidong Du, Jie Yan, Xing Hu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-04440.txt.
Description
Yutong Wu and ten co-authors teach language models to translate everyday math statements into formal, machine-checkable language. Autoformalization is a bottleneck for computer-verified mathematics: humans state theorems in words, proof assistants demand exact syntax. Their 2025 system, StepFun-Formalizer, targets two abilities the team finds wanting in current models. One is deep knowledge of the formal language itself, its objects and idioms. The other is reasoning: understanding the informal problem and aligning it precisely with the formal version.
Fusing knowledge with reasoning is the paper's recipe, against methods that lean on one alone. Low accuracy has kept autoformalization a demo rather than a tool. If models can formalize reliably, vast libraries of informal math become checkable. The authors frame the two abilities as separately trainable. Their work serves the growing formal mathematics community around proof assistants.
Catalog record
| CREATOR | Yutong Wu, Di Huang, Ruosi Wan, Yue Peng, Shijie Shang, Chenrui Cao, Lei Qi, Rui Zhang, Zidong Du, Jie Yan, Xing Hu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 79 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-04440