大模型明知问题无解却仍强行回答，根源是拒绝机制没连上识别信号。
Recognition-Refusal Misalignment in LLMs: Why Models Answer Structurally Unanswerable Questions
- 在隐藏状态中发现一个线性方向能区分可答与不可答问题。
- 识别无解的能力存在，但与拒绝回答的路径几乎垂直不匹配。
- 该现象在1.7B到70B参数模型中均出现，说明是路由而非编码缺陷。
大型语言模型常对结构上无解的问题（如计算cot(-540°)或判断(1).startswith("1")）作出回答，而非选择回避。我们探究这种失败是源于缺乏识别能力，还是识别后未正确引导至回避。在1.7B至70B参数的指令微调模型中，隐藏状态内存在单一线性方向，可有效分离可答与结构上不可能的数学和代码提示，表明模型在生成前已具备不可行性表征。然而，这一识别方向与主导有害内容拒绝的典型安全拒答方向几乎正交。基于领域行为定义的无效性感知方向虽更接近识别方向，但仍部分偏离，且与安全拒答方向保持近似正交。沿识别方向在生成时进行引导，可双向、剂量响应地改变结构化数学与代码任务的无效性感知行为，而随机方向则无效。基线/指令对比进一步表明，低余弦几何关系在预训练终点即已存在。因此，模型自信回答无解问题的现象更应归因于路由失败，而非编码缺失：模型拥有可用的“无合法答案”信号，但安全拒答路径未被对齐使用。
原文摘要 · Abstract (English)
Large language models often answer structurally unanswerable questions, such as computing cot(-540°) or evaluating (1).startswith("1"), instead of abstaining. We ask whether this failure reflects missing recognition or failed routing from recognition to abstention. Across instruction-tuned models from 1.7B to 70B parameters, a single linear direction in the hidden state separates answerable from structurally impossible math and code prompts, showing that models represent impossibility before generation. Yet this recognition direction is nearly orthogonal to the canonical safety-refusal direction that mediates trained harmful-content refusal. An in-domain behavior-defined invalidity-aware direction is closer to recognition, but only partially aligned with it, and remains near-orthogonal to safety refusal. Generation-time steering along the recognition direction changes invalidity-aware behavior bidirectionally and dose-responsively on structural math and code cells, while random directions do not. Base/instruct comparisons further show that the low-cosine geometry is already present at the pretraining endpoint. The confident-on-impossible failure is therefore better explained as a routing failure than as an encoding failure: the model has a usable "no admissible answer" signal, but the safety-refusal pathway is not aligned to use it.
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。