𝗥𝗔𝗚𝗔𝗦 — 𝗥𝗲𝘁𝗿𝗶𝗲𝘃𝗮𝗹 𝗔𝘂𝗴𝗺𝗲𝗻𝘁𝗲𝗱 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 𝗔𝘀𝘀𝗲𝘀𝘀𝗺𝗲𝗻𝘁

Published 2026-09-21 in AI Agents

Building a RAG application is one thing. But how do we know whether it is actually giving good answers? 🤔 That’s where 𝗥𝗔𝗚𝗔𝗦 comes in. 𝗥𝗔𝗚𝗔𝗦 is an open-source framework used to evaluate 𝗥𝗔𝗚 and 𝗟𝗟𝗠 applications. It helps us measure the quality of retrieved content and generated answers. 🔍 𝗦𝗼𝗺𝗲 𝗶𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝘁 𝗺𝗲𝘁𝗿𝗶𝗰𝘀: ✅ 𝗙𝗮𝗶𝘁𝗵𝗳𝘂𝗹𝗻𝗲𝘀𝘀 Is the answer supported by the retrieved context? ✅ 𝗔𝗻𝘀𝘄𝗲𝗿 𝗥𝗲𝗹𝗲𝘃𝗮𝗻𝗰𝘆 Does the answer actually address the user's question? ✅ 𝗖𝗼𝗻𝘁𝗲𝘅𝘁 𝗣𝗿𝗲𝗰𝗶𝘀𝗶𝗼𝗻 Are the relevant chunks ranked higher? ✅ 𝗖𝗼𝗻𝘁𝗲𝘅𝘁 𝗥𝗲𝗰𝗮𝗹𝗹 Did the system retrieve the information needed to answer the question? A simple way to remember it: 𝗥𝗔𝗚 → Retrieves information → Generates an answer 𝗥𝗔𝗚𝗔𝗦 → Evaluates how well it worked This is why evaluation is an important part of building reliable GenAI applications. Instead of just asking: ❌ “Does my RAG chatbot work?” We can ask: 👉 “How well is my RAG system performing?” 𝗕𝘂𝗶𝗹𝗱 → 𝗘𝘃𝗮𝗹𝘂𝗮𝘁𝗲 → 𝗜𝗺𝗽𝗿𝗼𝘃𝗲 🔄

More AI Agents articles · All collections · Practice challenges