RAGWarrant：在质量、成本、延迟和风险约束下的RAG策略推广的保留证据治理
RAGWarrant: Evidence-Preserving Governance for RAG Policy Promotion Under Quality, Cost, Latency, and Risk Constraints
AI总结：
RAGWarrant提出一个开源推广控制框架，将RAG策略部署视为受质量、成本、延迟和风险约束的证据决策，通过规范化评估、门控和可审计决策实现安全发布。
AI中文摘要：
检索增强生成系统广泛配备了指标、基准、追踪和自动评判器，但这些工具并不决定提议的策略变更是否安全发布。我们提出RAGWarrant，一个开源的推广控制框架，将部署视为受约束的证据决策而非排行榜选择。RAGWarrant规范化评估器输出和操作遥测，应用预先声明的质量和硬风险门控，分配证据类声明上限，保留负面结果，并发出可审计的PROMOTE、BLOCK、REJECT或INCONCLUSIVE决策。我们在T2-RAGBench、MultiHop-RAG、CRAG、HotpotQA、合成复现和有界本地生成实验上评估该框架。在HotpotQA上，由于答案质量超出声明余量，操作节省被阻止。一个有界的CRAG研究选择了成本较低且与质量挂钩的策略，但相关的生成收益不稳定，且一个保留的防护栏未能关闭。我们声称提供可审计的推广控制抽象，而非优化器优越性、人工验证或生产就绪性。标记的工件可从全新克隆中复现，作为加固的Docker作业运行，接受外部评估器导出，并验证工件完整性。
英文摘要：
Retrieval-augmented generation systems are extensively instrumented with metrics, benchmarks, traces, and automated judges, but these tools do not decide whether a proposed policy change is safe to release. We present RAGWarrant, an open-source promotion-control framework that treats deployment as a constrained evidence decision rather than a leaderboard choice. RAGWarrant normalizes evaluator outputs and operational telemetry, applies predeclared quality and hard-risk gates, assigns evidence-class claim ceilings, preserves negative outcomes, and emits auditable PROMOTE, BLOCK, REJECT, or INCONCLUSIVE decisions. We evaluate the framework across T2-RAGBench, MultiHop-RAG, CRAG, HotpotQA, synthetic reproduction, and bounded local generative experiments. On HotpotQA, operational savings were blocked because answer quality fell beyond the declared margin. A bounded CRAG study selected a lower-cost quality-tied policy, but related generative gains were unstable and a held-out guardrail failed closed. We claim an auditable promotion-control abstraction, not optimizer superiority, human validation, or production readiness. The tagged artifact reproduces from a fresh clone, runs as a hardened Docker job, accepts external evaluator exports, and verifies artifact integrity.
