正确思考，而非更多思考：用于数值声明验证的测试时扩展
Think Right, Not More: Test-Time Scaling for Numerical Claim Verification
AI总结：
本文针对复杂数值声明核查，提出利用 TTS 生成多条推理路径并由 VERIFIERFC 选择正确判定，还通过复杂度感知的自适应机制将效率提升至标准 TTS 的 1.8 倍，性能较单轮方法提高 18.8%。
摘要：
核查现实世界中的声明，尤其是数值声明，本身就很复杂，需要多步推理和数值推理来验证声明的各个方面。尽管包括推理模型在内的大型语言模型（LLM）已经取得了巨大进展，但它们在核查需要组合推理与数值推理相结合的现实世界声明方面仍有不足。它们无法理解数值方面的细微差别，并且容易受到推理漂移问题的影响，即模型无法将多样信息置于上下文中，从而导致推理过程的误解和回溯。在本工作中，我们系统地探索了在核查复杂数值声明任务上为 LLM 扩展测试时计算（TTS），这需要从 LLM 引出多条推理路径。我们训练了一个验证器模型（VERIFIERFC），以在可能推理路径构成的空间中导航，并选择一条能够得出正确判定的路径。我们观察到，TTS 有助于缓解推理漂移问题，从而为数值声明核查带来显著的性能提升。为了提高 TTS 中的计算效率，我们引入了一种自适应机制，该机制根据所感知到的声明复杂性有选择地执行 TTS。该方法实现了比标准 TTS 高 1.8 倍的效率，同时相较于单轮声明验证方法带来了显著的 18.8% 性能提升。我们的代码和数据可在 https://github.com/VenkteshV/VerifierFC 获取。
英文摘要：
Fact-checking real-world claims, particularly numerical claims, is inherently complex that require multistep reasoning and numerical reasoning for verifying diverse aspects of the claim. Although large language models (LLMs) including reasoning models have made tremendous advances, they still fall short on fact-checking real-world claims that require a combination of compositional and numerical reasoning. They are unable to understand nuance of numerical aspects, and are also susceptible to the reasoning drift issue, where the model is unable to contextualize diverse information resulting in misinterpretation and backtracking of reasoning process. In this work, we systematically explore scaling test-time compute (TTS) for LLMs on the task of fact-checking complex numerical claims, which entails eliciting multiple reasoning paths from an LLM. We train a verifier model (VERIFIERFC) to navigate this space of possible reasoning paths and select one that could lead to the correct verdict. We observe that TTS helps mitigate the reasoning drift issue, leading to significant performance gains for fact-checking numerical claims. To improve compute efficiency in TTS, we introduce an adaptive mechanism that performs TTS selectively based on the perceived complexity of the claim. This approach achieves 1.8x higher efficiency than standard TTS, while delivering a notable 18.8% performance improvement over single-shot claim verification methods. Our code and data can be found at https://github.com/VenkteshV/VerifierFC
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
