📄 中文摘要
大型语言模型虽然强大，但仍然面临一个主要问题——幻觉。在构建 AI 助手的过程中，常常发现模型能够生成听起来令人信服的答案，但实际上并没有基于真实数据。这促使对检索增强生成（RAG）进行探索，构建一个系统，使 Gemini 能够使用真实文档回答问题，而不是依赖猜测。该系统通过整合真实数据，显著提高了回答的准确性和可靠性，解决了大型语言模型的局限性。
从幻觉到扎根的人工智能：构建一个基于 Qdrant 的 Gemini RAG 系统
📄 English Summary
From Hallucinations to Grounded AI: Building a Gemini RAG System with Qdrant
Large Language Models (LLMs) are powerful but face a significant challenge known as hallucinations. In the process of developing AI assistants, it became evident that these models could produce answers that sounded plausible yet were not grounded in actual data. This realization led to an exploration of Retrieval-Augmented Generation (RAG) to create a system that enables Gemini to respond to queries using real documents instead of mere guesses. By integrating authentic data, this system significantly enhances the accuracy and reliability of responses, addressing the limitations of LLMs.