增加推理计算能提升模型鲁棒性，但前提是中间推理步骤不被攻击者看到。
Does More Inference-Time Compute Really Help Robustness?
- 用预算强制策略让小模型也受益于推理时扩容。
- 当推理过程可见时，算力越多反而越不安全，出现反向缩放现象。
- 适用于需防范推理泄露的高风险场景，如工具调用模型。
近期研究发现，大型专有推理模型通过增加推理时计算可提升鲁棒性。本文首先表明，小型开源模型（如 DeepSeek R1、Qwen3、Phi-reasoning）也可通过简单的预算强制策略实现推理时扩展带来的收益。更重要的是，我们揭示并批判性检验了此前工作的隐含假设：中间推理步骤对攻击者是隐藏的。当该假设被打破，即推理过程可被直接访问时，我们发现一个直观且经实证验证的反向缩放规律——随着推理时计算资源增加，模型鲁棒性反而持续下降。最后，我们讨论了在实际场景中，即使推理链被隐藏，模型仍可能因工具集成推理或高级推理提取攻击而存在漏洞。综合来看，推理时扩展的鲁棒性优势高度依赖对抗环境和部署上下文。我们呼吁实践者在安全敏感的真实应用中，谨慎权衡这一微妙权衡。
原文摘要 · Abstract (English)
Recently, Zaremba et al. demonstrated that increasing inference-time computation improves robustness in large proprietary reasoning LLMs. In this paper, we first show that smaller-scale, open-source models (e.g., DeepSeek R1, Qwen3, Phi-reasoning) can also benefit from inference-time scaling using a simple budget forcing strategy. More importantly, we reveal and critically examine an implicit assumption in prior work: intermediate reasoning steps are hidden from adversaries. By relaxing this assumption, we identify an important security risk, intuitively motivated and empirically verified as an inverse scaling law: if intermediate reasoning steps become explicitly accessible, increased inference-time computation consistently reduces model robustness. Finally, we discuss practical scenarios where models with hidden reasoning chains are still vulnerable to attacks, such as models with tool-integrated reasoning and advanced reasoning extraction attacks. Our findings collectively demonstrate that the robustness benefits of inference-time scaling depend heavily on the adversarial setting and deployment context. We urge practitioners to carefully weigh these subtle trade-offs before applying inference-time scaling in security-sensitive, real-world applications.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。