Efficient PRM Training Data Synthesis via Formal Verification
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_ | 1866910113607974912 |
|---|---|
| author | Kamoi, Ryo Zhang, Yusen Zhang, Nan Das, Sarkar Snigdha Sarathi Zhang, Ranran Haoran Yin, Wenpeng Zhang, Rui |
| author_facet | Kamoi, Ryo Zhang, Yusen Zhang, Nan Das, Sarkar Snigdha Sarathi Zhang, Ranran Haoran Yin, Wenpeng Zhang, Rui |
| contents | Process Reward Models (PRMs) have emerged as a promising approach for improving LLM reasoning capabilities by providing process supervision over reasoning traces. However, existing approaches for constructing PRM training data remain costly and noisy, as they typically rely on human annotation or sampling-based labeling methods that require repeated LLM calls. In this work, we propose FoVer, a framework that synthesizes PRM training data from formal reasoning tasks by annotating step-level error labels using formal verification tools such as Z3 and Isabelle. By leveraging formal verification, FoVer enables efficient and accurate PRM data construction without requiring human annotation or additional LLM calls. Using FoVer, we create PRM training data from formal logic and theorem proving tasks. Experiments on 12 reasoning benchmarks show that fine-tuning on our training data improves PRMs not only on math and logic reasoning tasks, which are informal variants of the training tasks, but also on NLI and BBH benchmarks, which differ substantially from the tasks used to construct the training data. These results demonstrate the practical effectiveness of FoVer, showing that PRM training data created using formal verification improves PRMs on informal reasoning tasks written in natural language. The datasets, models, and code are provided at https://github.com/psunlpgroup/FoVer. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_15960 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Efficient PRM Training Data Synthesis via Formal Verification Kamoi, Ryo Zhang, Yusen Zhang, Nan Das, Sarkar Snigdha Sarathi Zhang, Ranran Haoran Yin, Wenpeng Zhang, Rui Computation and Language Process Reward Models (PRMs) have emerged as a promising approach for improving LLM reasoning capabilities by providing process supervision over reasoning traces. However, existing approaches for constructing PRM training data remain costly and noisy, as they typically rely on human annotation or sampling-based labeling methods that require repeated LLM calls. In this work, we propose FoVer, a framework that synthesizes PRM training data from formal reasoning tasks by annotating step-level error labels using formal verification tools such as Z3 and Isabelle. By leveraging formal verification, FoVer enables efficient and accurate PRM data construction without requiring human annotation or additional LLM calls. Using FoVer, we create PRM training data from formal logic and theorem proving tasks. Experiments on 12 reasoning benchmarks show that fine-tuning on our training data improves PRMs not only on math and logic reasoning tasks, which are informal variants of the training tasks, but also on NLI and BBH benchmarks, which differ substantially from the tasks used to construct the training data. These results demonstrate the practical effectiveness of FoVer, showing that PRM training data created using formal verification improves PRMs on informal reasoning tasks written in natural language. The datasets, models, and code are provided at https://github.com/psunlpgroup/FoVer. |
| title | Efficient PRM Training Data Synthesis via Formal Verification |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2505.15960 |