arXiv cs.AI· Yichi Zhang, Zhenxiang Huang, Yuyang Zhang, Yuchen Yang·· 3 天前AI 评分24
SCRouter：面向安全的大语言模型路由如何按查询选择模型与预算
Rethinking Quality in Large Language Model Routing: Security-Conditioned Model-and-Budget Selection
AI 导读
针对现有 LLM 路由器只用单一质量指标、难以兼顾普通查询与安全敏感查询的问题，研究者提出 SCRouter，分别预测响应的效用与安全性，并用一个小型探针为每个查询确定两者权重，再结合成本选择模型与 token 预算。
正文 · 原文
Abstract:Large language model (LLM) routers reduce serving cost by choosing which model to call and how many output tokens to grant it for each query. Existing routers rank model-budget actions using a single quality metric. This does not work for mixed workloads. For ordinary queries, a good answer is correct and useful (utility); for security-sensitive queries, a good answer refuses harmful requests or resists malicious instructions (security). These two objectives prefer different models and budgets.
We propose SCRouter, a router that learns when to prioritize security and how to allocate the token budget. For each model-budget pair, it estimates the probability that the response finishes within budget. It uses this probability to predict utility and security separately. A small probe determines how to weight these two predictions for each query. SCRouter then selects the model and budget by balancing the weighted utility and security against cost. We also build SCRouterBench, a benchmark of 38,869 utility and security queries spanning nine security dimensions, each answered under all 144 combinations of 12 models and 12 budgets, plus 2,753 out-of-domain queries. SCRouter improves the baselines by 13-29 points in-domain and 19-23 points out-of-domain.
| Subjects: | Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR) |
| Cite as: | arXiv:2610.18566 [cs.AI] |
| (or arXiv:2610.18566v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.18566 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yichi Zhang [view email]
[v1]
Tue, 13 Oct 2026 15:36:57 UTC (412 KB)
来源：arXiv cs.AI · arxiv.org