Retrieval-Augmented Generation (RAG) systems often fail on multi-hop queries when the initial retrieval misses a bridge fact. Prior corrective approaches, such as Self-RAG, CRAG, and Adaptive-$k$, typically address this by \textit{adding} more context or pruning existing lists. However, simply expanding the context window often leads to \textbf{context dilution}, where distractors crowd out relevant information. We propose \textbf{SEAL-RAG}, a training-free controller that adopts a \textbf{``replace, don't expand''} strategy to fight context dilution under a fixed retrieval depth $k$. SEAL executes a (\textbf{S}earch $\rightarrow$ \textbf{E}xtract $\rightarrow$ \textbf{A}ssess $\rightarrow$ \textbf{L}oop) cycle: it performs on-the-fly, entity-anchored extraction to build a live \textit{gap specification} (missing entities/relations), triggers targeted micro-queries, and uses \textit{entity-first ranking} to actively swap out distractors for gap-closing evidence. We evaluate SEAL-RAG against faithful re-implementations of Basic RAG, CRAG, Self-RAG, and Adaptive-$k$ in a shared environment on \textbf{HotpotQA} and \textbf{2WikiMultiHopQA}. On HotpotQA ($k=3$), SEAL improves answer correctness by \textbf{+3--13 pp} and evidence precision by \textbf{+12--18 pp} over Self-RAG. On 2WikiMultiHopQA ($k=5$), it outperforms Adaptive-$k$ by \textbf{+8.0 pp} in accuracy and maintains \textbf{96\%} evidence precision compared to 22\% for CRAG. These gains are statistically significant ($p<0.001$). By enforcing fixed-$k$ replacement, SEAL yields a predictable cost profile while ensuring the top-$k$ slots are optimized for precision rather than mere breadth. We release our code and data at https://github.com/mosherino/SEAL-RAG.

翻译：检索增强生成（RAG）系统在处理多跳查询时，若初始检索遗漏了桥梁事实，往往会失败。现有的纠正方法，如Self-RAG、CRAG和Adaptive-$k$，通常通过*添加*更多上下文或修剪现有列表来解决此问题。然而，单纯扩展上下文窗口常导致**上下文稀释**，即干扰信息挤占了相关信息。我们提出**SEAL-RAG**，一种免训练的控制器，采用**“替换，而非扩展”**策略，在固定检索深度$k$下对抗上下文稀释。SEAL执行一个（**S**earch → **E**xtract → **A**ssess → **L**oop）循环：它执行动态的、以实体为锚点的提取，以构建实时*缺口规范*（缺失的实体/关系），触发有针对性的微查询，并使用*实体优先排序*主动将干扰信息替换为填补缺口的证据。我们在共享环境中，于**HotpotQA**和**2WikiMultiHopQA**数据集上，将SEAL-RAG与Basic RAG、CRAG、Self-RAG和Adaptive-$k$的忠实复现版本进行了评估。在HotpotQA（$k=3$）上，SEAL将答案正确率较Self-RAG提升了**+3--13个百分点**，证据精确率提升了**+12--18个百分点**。在2WikiMultiHopQA（$k=5$）上，其准确率较Adaptive-$k$高出**+8.0个百分点**，并保持了**96%**的证据精确率，而CRAG仅为22%。这些提升具有统计显著性（$p<0.001$）。通过强制执行固定$k$替换，SEAL产生了可预测的成本特征，同时确保前$k$个槽位为精确度而非单纯广度进行了优化。我们在https://github.com/mosherino/SEAL-RAG发布了代码和数据。