面试题库

全部 Django Elasticsearch FastAPI Flask Function Call JVM Java IO Java 后端 Java 基础 Java 并发 Java 新特性 Java 集合 JavaWeb Kafka LLM LangChain LangGraph MongoDB MyBatis MySQL Netty Nginx Python RAG RabbitMQ Redis RocketMQ Spring Spring AI SpringBoot SpringCloud SpringMVC ZooKeeper 分布式 后端技术 大厂面经 大数据 微服务 提示词工程 操作系统 数据库 数据结构 智能体 海康威视2026秋季社招专项 消息队列 监控运维 算法 系统设计 缓存 计算机基础 计算机网络 设计模式 面试题 高可用 高并发
LLM 中等

What are the different types of LLM fine-tuning?

作答 →
LLM 中等

What are the different phases in LLM development?

作答 →
LLM 中等

How does batching generally help with LLM inference efficiency?

作答 →
LLM 中等

What is batch inference, and how does it differ from single-query inference?

作答 →
LLM 中等

What is latency in LLM inference, and why is it important?

作答 →
RAG 中等

Why is it important for RAG systems to optimize both context precision and context recall simultaneously? What trade-offs might occur?

作答 →
RAG 中等

If your RAG retriever consistently shows Context Recall scores below 0.6, what are the three potential root causes?

作答 →
RAG 中等

Describe a scenario where a RAG system might achieve high Context Recall but still produce poor answers. What complementary metrics would you use alongside Context Recall to get a complete picture of

作答 →
RAG 中等

Why is MRR unsuitable when there are multiple relevant chunks per query, and how does MAP address this limitation?

作答 →
RAG 中等

Explain the difference between Precision@k and Recall@k in the context of RAG. When might you prefer one over the other?

作答 →
RAG 中等

How would you evaluate the effectiveness of a reranker in a RAG system? Which metrics (e.g., MRR, MAP, NDCG) would you prioritize and why?

作答 →
RAG 中等

A RAG system retrieves 20 candidate document chunks but can only fit 5 in the LLM's context window. Without re-ranking, how might this limitation affect response quality, and what specific problems wo

作答 →