CoSineVerifier: Tool-Augmented Answer Verification for Computation-Oriented Scientific Questions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Feng, Ruixiang, An, Zhenwei, Wen, Yuntao, Le, Ran, Jia, Yiming, Yang, Chen, Chen, Zongchao, Chen, Lisi, Gao, Shen, Shang, Shuo, Song, Yang, Zhang, Tao
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911296210862080
author Feng, Ruixiang
An, Zhenwei
Wen, Yuntao
Le, Ran
Jia, Yiming
Yang, Chen
Chen, Zongchao
Chen, Lisi
Gao, Shen
Shang, Shuo
Song, Yang
Zhang, Tao
author_facet Feng, Ruixiang
An, Zhenwei
Wen, Yuntao
Le, Ran
Jia, Yiming
Yang, Chen
Chen, Zongchao
Chen, Lisi
Gao, Shen
Shang, Shuo
Song, Yang
Zhang, Tao
contents Answer verification methods are widely employed in language model training pipelines spanning data curation, evaluation, and reinforcement learning with verifiable rewards (RLVR). While prior work focus on developing unified verifiers applicable across multiple reasoning scenarios, significant challenges remain in computation-oriented scientific domains, such as algebraic equivalence checking and physical constant substitution. In this paper, we introduce \model, a tool-augmented verifier that leverages external executors to perform precise computations and symbolic simplifications. \model enables robust verification that goes beyond simple semantic matching. We propose a novel two-stage pipeline, which begin with cold-start fine-tuning and followed by multi-turn reinforcement learning with tool integration. Extensive experiments conducted on STEM subjects, general QA, and long-form reasoning tasks demonstrates strong generalization of \model. The results shows that the \model achieves state-of-the-art performance on VerifyBench-Hard and SCI-Bench. And we also employ our \model in RLVR as a reward model, the results show that it consistently outperforms both rubric-based and model-based verifiers on AIME'24 and AIME'25, demonstrating strong potential to enhance reasoning capabilities of LLM. Our model is released at \hyperlink{https://huggingface.co/Nanbeige/CoSineVerifier-Tool-4B}{https://huggingface.co/Nanbeige/CoSineVerifier-Tool-4B}.
format Preprint
id arxiv_https___arxiv_org_abs_2512_01224
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CoSineVerifier: Tool-Augmented Answer Verification for Computation-Oriented Scientific Questions
Feng, Ruixiang
An, Zhenwei
Wen, Yuntao
Le, Ran
Jia, Yiming
Yang, Chen
Chen, Zongchao
Chen, Lisi
Gao, Shen
Shang, Shuo
Song, Yang
Zhang, Tao
Machine Learning
Answer verification methods are widely employed in language model training pipelines spanning data curation, evaluation, and reinforcement learning with verifiable rewards (RLVR). While prior work focus on developing unified verifiers applicable across multiple reasoning scenarios, significant challenges remain in computation-oriented scientific domains, such as algebraic equivalence checking and physical constant substitution. In this paper, we introduce \model, a tool-augmented verifier that leverages external executors to perform precise computations and symbolic simplifications. \model enables robust verification that goes beyond simple semantic matching. We propose a novel two-stage pipeline, which begin with cold-start fine-tuning and followed by multi-turn reinforcement learning with tool integration. Extensive experiments conducted on STEM subjects, general QA, and long-form reasoning tasks demonstrates strong generalization of \model. The results shows that the \model achieves state-of-the-art performance on VerifyBench-Hard and SCI-Bench. And we also employ our \model in RLVR as a reward model, the results show that it consistently outperforms both rubric-based and model-based verifiers on AIME'24 and AIME'25, demonstrating strong potential to enhance reasoning capabilities of LLM. Our model is released at \hyperlink{https://huggingface.co/Nanbeige/CoSineVerifier-Tool-4B}{https://huggingface.co/Nanbeige/CoSineVerifier-Tool-4B}.
title CoSineVerifier: Tool-Augmented Answer Verification for Computation-Oriented Scientific Questions
topic Machine Learning
url https://arxiv.org/abs/2512.01224