为RAG生成设计可满足SLO的动态决策机制，避免盲目追求低成本。
SLO-Conditioned Action Routing for Retrieval-Augmented Generation: Objective Ablation and Failure Modes
- 基于查询选择检索深度与生成模式，实现每请求的精细控制。
- 在质量优先时，学习策略比固定基线节省成本；在廉价优先时易出现拒绝崩溃。
- 强调失败模式分析，适合关注系统稳定性的工程师参考。
检索增强生成（RAG）面临实际控制难题：需针对每个查询选择检索深度与生成行为以满足服务等级目标（SLO），如成本、拒答率和幻觉风险。本文将每查询控制建模为一个离散动作：选择检索深度、生成模式（防护型或自动型）或直接拒绝。通过在SQuAD 2.0上执行各动作并记录准确率、令牌成本、幻觉/拒答指标及加权奖励，构建离线日志数据集。评估两种简单策略学习目标：基于状态最优动作的监督分类（Argmax-CE）及其奖励加权变体（Argmax-CE-WT）。在多数场景中，固定基线（低k值+防护提示）表现良好；学习策略主要在质量导向的SLO下带来额外成本节省，但在成本导向的SLO下因拒答被重奖而可能引发拒绝崩溃。贡献在于提供一个可复现的RAG系统中面向SLO的控制案例研究，重点在于失败模式识别与报告规范，而非提出新检索器或语言模型。
原文摘要 · Abstract (English)
Retrieval-augmented generation (RAG) introduces a practical control problem: retrieval depth and generation behavior must be chosen per query to satisfy service-level objectives (SLOs) such as cost, refusal rate, and hallucination risk. This work models per-query control as a small discrete action: choose a retrieval depth and a generation mode (guarded vs. auto), or refuse. An offline logged dataset is constructed from SQuAD 2.0 by executing each action and recording accuracy, token cost, hallucination/refusal indicators, and an SLO-weighted reward. Two simple policy-learning objectives are evaluated: supervised classification of the per-state best action (Argmax-CE) and a reward-weighted variant (Argmax-CE-WT). Across the evaluated settings, a strong fixed baseline (low k, guarded prompting) performs competitively; learned policies mainly provide additional cost savings under a quality-focused SLO and can exhibit refusal collapse under a cheap SLO when refusal is heavily rewarded. The contribution is a reproducible case study of SLO-aware control for RAG pipelines, emphasizing failure modes and reporting conventions rather than proposing a new retriever or language model.
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。