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 role does alignment tuning play in improving an LLM's usability?

作答 →
LLM 中等

Explain the reason behind the effectiveness of Chain-of-Thought (CoT) prompting

作答 →
LLM 中等

What is Chain-of-Thought (CoT) prompting, and when is it most useful?

作答 →
LLM 中等

What are the possible options for accelerating LLM inference?

作答 →
LLM 中等

What is mixed precision (e.g., FP16) and why is it used during inference?

作答 →
LLM 中等

What is continuous batching, and how does it differ from static batching?

作答 →
LLM 中等

Explain the role of flash attention in reducing memory bottlenecks during inference

作答 →
LLM 中等

Explain the role of decoding strategy in LLM text generation

作答 →
LLM 中等

How can techniques like mixture-of-experts (MoE) optimize inference efficiency?

作答 →
LLM 中等

Explain the trade-offs between batching and latency in LLM serving

作答 →
LLM 中等

What is the softmax function, and where is it applied in the Transformer model?

作答 →
LLM 中等

Explain why self-attention in the decoder is referred to as cross-attention. How does it differ from self-attention in the encoder?

作答 →