RAG on Mengboy 技术笔记

RAG Accuracy Playbook: Retrieval Recall, Re-Ranking, and Evaluation Loop

Tue, 17 Feb 2026 10:56:00 +0800

If your RAG system feels unreliable, switching to a more expensive LLM is usually the wrong first move. In most cases, the bottleneck is retrieval quality: weak recall, poor ranking, and no measurement loop.

This guide gives a practical path: make recall broader, make ranking sharper, then close the loop with offline + online evaluation.

RAG 不准怎么办：检索召回、重排与评估闭环落地指南

Tue, 17 Feb 2026 10:56:00 +0800

很多团队做 RAG 的第一反应是“把 embedding 换成更贵的模型”，结果成本上去了，效果却不稳定。真正的问题通常不在生成，而在检索链路：召回不全、排序不准、评估缺失。

这篇给一套可直接落地的做法：先把召回做厚，再把重排做准，最后用离线 + 在线指标形成持续优化闭环。