为大模型服务设计可信赖的异常输入拒绝机制
SCOPE: Sequential Conformal Probing for Reliable OOD Rejection in LLM Services
- 从模型中间层提取边界信号，构建可解释的拒绝门控
- 在多个模型和边界条件下，拒识率显著优于传统方法
- 提供理论保证，适合对安全性要求高的部署场景
大语言模型服务中，拒绝超出定义分布范围的输入至关重要，未支持请求应在完整生成前被过滤。现有异常输入检测方法多依赖最终输出或最后一层表征，难以定位模型内部服务边界信号的清晰编码位置，且缺乏对保留输入的理论保障。本文提出SCOPE（序列置信探测与评估），通过选择可读取的隐藏层，构建基于归纳校准的置信门控，并利用超鞅e过程对持续的服务边界证据进行认证。在多个LLM主干模型和六种精心设计的边界条件下实验表明，SCOPE在门控级拒绝性能上优于标准最终层检测器，同时揭示了不同异常边界在隐空间中呈现不同的几何形态。
原文摘要 · Abstract (English)
Rejecting inputs outside the defined in-distribution (IND) service scope is critical for large language model (LLM) services, where unsupported requests should be filtered before full generation. Existing out-of-distribution (OOD) detectors often rely on final outputs or final-layer representations, leaving unclear where service-boundary signals are most clearly encoded inside the model; they also lack a theoretical guarantee for held-out inputs. In this paper, we introduce SCOPE (Sequential Conformal OOD Probing and Evaluation), a framework that selects a readable hidden layer, constructs a conformal gate with IND calibration, and uses a supermartingale e-process to certify persistent service-boundary evidence. Experiments across multiple LLM backbones and six carefully designed boundary conditions show that SCOPE improves gate-level rejection over standard final-layer detectors, while revealing how different OOD boundaries take different geometric forms in hidden space.
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。