CASE EVIDENCE / A RECORD
Vidhyadhara Rao Kotagiri / KL University RAG-based Assistant 公开的知识库与检索问答案例，来源为 公开代码库，复核于 2026-06-27T06:41:19Z。
原始记录：KL University RAG-based Assistant uses Qwen2.5-72B-Instruct for campus Q&A
Chinese Brief
Vidhyadhara Rao Kotagiri / KL University RAG-based Assistant 公开的知识库与检索问答案例，来源为 公开代码库，复核于 2026-06-27T06:41:19Z。 Model Atlas 将它标记为 A 类证据，因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。 Model Atlas 不把 benchmark、教程、发布说明或集合页包装成真实案例。
任务
这是一个围绕知识检索和问答的真实任务，公开材料可以回溯到具体使用者和具体产物。 原始资料写作：A Streamlit RAG assistant for answering questions related to KL University: it embeds user queries, retrieves relevant chunks from a local text corpus, and sends the retrieved context to Qwen/Qwen2.5-72B-Instruct via th…
公开材料提供公开代码、README 或项目配置，可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作：The application returns campus-policy answers in a web UI; the README example asks about KL University dress-code policy and describes retrieving the top two similar chunks before displaying the generated answer.
Qwen2.5 72B 在该案例中承担知识检索和问答相关的生成、分析、编排或实现角色。 原始资料写作：Qwen2.5-72B-Instruct is the text-generation component that turns retrieved KL University context into the final natural-language answer.
当前判断基于公开材料；若产物下线、仓库变更或模型参与比例仅来自作者自述，需要在引用前重新复核。 原始资料写作：Public repo evidence binds the exact Hugging Face model id; artifact is the source repository rather than a hosted production deployment.