不改模型参数，用向量引导大模型高效推理。
Model Whisper: Steering Vectors Unlock Large Language Models' Potential in Test-time
- 输入前加轻量向量，冻结模型参数优化它。
- 数学任务上性能提升超45%，且跨任务通用性强。
- 适合需要快速适配新任务的场景，无需调参。
如何高效激活大语言模型在特定任务或新分布下的推理潜力，是当前关键挑战。现有测试时适配方法通常需调整模型参数，不仅计算开销大，还可能损害模型原有能力。为此，我们提出一种轻量级组件——测试时引导向量（TTSV），将其前置到输入中，同时保持大模型参数完全冻结。通过在测试数据上优化TTSV以最小化模型输出熵，使模型进入更高置信度的内部状态，从而激活其与当前任务最相关的内在能力。TTSV优化轻量高效，可作为即插即用的增强模块。大量实验验证了该方法在基础模型和推理增强模型上的有效性。例如，在MATH500任务上，对Qwen2.5-Math-7B模型实现45.88%的相对性能提升，对Qwen3-4B模型达16.22%。此外，该方法表现出强泛化能力，其引导向量在多种任务间具有高度可迁移性。
原文摘要 · Abstract (English)
It is a critical challenge to efficiently unlock the powerful reasoning potential of Large Language Models (LLMs) for specific tasks or new distributions. Existing test-time adaptation methods often require tuning model parameters, which is not only computationally expensive but also risks degrading the model's pre-existing abilities.To address this, we introduce a lightweight component, Test-Time Steering Vectors (TTSV), which is prepended to the input while keeping the LLM's parameters entirely frozen. By optimizing the TTSV on test data to minimize the model's output entropy, we steer the model towards an internal state of higher confidence, activating its inherent abilities most relevant to the current task. TTSV is both lightweight and highly efficient to optimize, making it a true plug-and-play enhancement. Extensive experiments validate our approach's effectiveness on both base models and reasoning-enhanced models. For instance, on the MATH500 task, TTSV achieves a 45.88% relative performance gain on the Qwen2.5-Math-7B model and a 16.22% relative gain on the Qwen3-4B model. Furthermore, our approach exhibits robust generalization, with its steering vectors proving highly transferable across diverse tasks.
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。