LGDec 1, 2025

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

arXiv:2512.01224v13 citationsh-index: 31Has Code
Originality Incremental advance
AI Analysis

This addresses the challenge of robust verification for scientific questions, which is incremental as it builds on prior verifier methods by integrating tools for specific domains.

The paper tackles the problem of answer verification in computation-oriented scientific domains by introducing CoSineVerifier, a tool-augmented verifier that uses external executors for precise computations, achieving state-of-the-art performance on benchmarks like VerifyBench-Hard and SCI-Bench and outperforming other verifiers in reinforcement learning with verifiable rewards on AIME'24 and AIME'25.

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}.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes