https://guijiagi.com/tags/%E6%8E%A8%E7%90%86/Recent content in 推理 on 硅基 AGI · 智能体学习与测评Hugozh-cn本站内容采用 CC BY-NC-SA 4.0 国际许可协议授权Thu, 16 Jul 2026 11:37:00 +0800https://guijiagi.com/posts/b1-acff1ec2/Thu, 16 Jul 2026 11:37:00 +0800https://guijiagi.com/posts/b1-acff1ec2/系统梳理思维链推理技术的发展，从基础CoT到Tree of Thoughts再到推理时Scalinghttps://guijiagi.com/posts/article-18/Sun, 12 Jul 2026 19:50:00 +0800https://guijiagi.com/posts/article-18/从CoT到ToT再到GoT，系统梳理Prompt工程中思维结构化的技术演进和实践方法https://guijiagi.com/posts/b2-e84f5637/Sun, 12 Jul 2026 19:50:00 +0800https://guijiagi.com/posts/b2-e84f5637/从CoT到ToT再到GoT，系统梳理Prompt工程中思维结构化的技术演进和实践方法https://guijiagi.com/posts/chain-of-thought-2026-advanced/Thu, 02 Jul 2026 11:05:00 +0800https://guijiagi.com/posts/chain-of-thought-2026-advanced/2026年思维链（Chain-of-Thought）提示工程进阶技巧与最佳实践https://guijiagi.com/posts/neuro-symbolic-ai-2026/Thu, 02 Jul 2026 10:38:00 +0800https://guijiagi.com/posts/neuro-symbolic-ai-2026/神经网络与符号AI的融合：2026年的最新进展https://guijiagi.com/posts/chain-of-thought-advanced-tot-got/Sun, 28 Jun 2026 10:05:00 +0800https://guijiagi.com/posts/chain-of-thought-advanced-tot-got/深入解析 CoT 的进化路线——Tree-of-Thought 与 Graph-of-Thought 推理框架，附代码实现与效果对比https://guijiagi.com/posts/chain-of-thought-advanced/Sat, 27 Jun 2026 15:00:00 +0800https://guijiagi.com/posts/chain-of-thought-advanced/从基础CoT到高级变体的全面指南，掌握思维链提示工程的进阶技巧https://guijiagi.com/posts/hermes-reasoning-capability/Sat, 27 Jun 2026 15:00:00 +0800https://guijiagi.com/posts/hermes-reasoning-capability/Hermes推理能力评测https://guijiagi.com/posts/kv-cache-optimization/Sat, 27 Jun 2026 15:00:00 +0800https://guijiagi.com/posts/kv-cache-optimization/KV Cache优化策略详解https://guijiagi.com/posts/sglang-inference-framework/Sat, 27 Jun 2026 15:00:00 +0800https://guijiagi.com/posts/sglang-inference-framework/SGLang推理框架解析https://guijiagi.com/posts/vllm-opensource-inference/Sat, 27 Jun 2026 15:00:00 +0800https://guijiagi.com/posts/vllm-opensource-inference/vLLM开源推理引擎https://guijiagi.com/posts/cot-variants-comparison/Sat, 27 Jun 2026 15:00:00 +0800https://guijiagi.com/posts/cot-variants-comparison/Zero-shot CoT、Few-shot CoT、Auto-CoT等思维链变体的深度对比与选型指南https://guijiagi.com/posts/inference-scaling-laws/Sat, 27 Jun 2026 15:00:00 +0800https://guijiagi.com/posts/inference-scaling-laws/推理时计算Scaling Lawshttps://guijiagi.com/posts/prompt-chain-design/Fri, 26 Jun 2026 14:00:00 +0800https://guijiagi.com/posts/prompt-chain-design/系统讲解 Prompt 链式设计的原理、模式与工程实践，构建从简单推理到复杂任务分解的完整方法论https://guijiagi.com/posts/react-vs-plan-execute/Fri, 26 Jun 2026 14:00:00 +0800https://guijiagi.com/posts/react-vs-plan-execute/深入对比 ReAct 与 Plan-and-Execute 两种主流智能体推理范式的工作原理、优劣势及适用场景，指导工程选型。https://guijiagi.com/posts/ai-inference-serving/Thu, 25 Jun 2026 12:00:00 +0800https://guijiagi.com/posts/ai-inference-serving/系统讲解 AI 推理服务的架构演进，涵盖模型加载、批处理、GPU 调度与弹性伸缩https://guijiagi.com/posts/chain-of-thought-mastery/Thu, 25 Jun 2026 12:00:00 +0800https://guijiagi.com/posts/chain-of-thought-mastery/深入解析 Chain of Thought 提示技术，从基础概念到高级变体，全面掌握 CoT 推理增强的实践方法。https://guijiagi.com/posts/sglang-inference/Thu, 25 Jun 2026 12:00:00 +0800https://guijiagi.com/posts/sglang-inference/深入解析 SGLang 推理引擎的核心技术原理、性能优势与生产部署实践，涵盖 RadixAttention、结构化生成、并发处理等关键特性。https://guijiagi.com/posts/vllm-production-guide/Thu, 25 Jun 2026 12:00:00 +0800https://guijiagi.com/posts/vllm-production-guide/从架构原理到生产部署，全面解析 vLLM 高吞吐推理引擎的核心技术、性能调优与运维方案。https://guijiagi.com/posts/continuous-batching/Thu, 25 Jun 2026 12:00:00 +0800https://guijiagi.com/posts/continuous-batching/深度解析 vLLM 连续批处理（Continuous Batching）如何通过动态请求调度和 PagedAttention 实现高吞吐 LLM 推理服务。https://guijiagi.com/posts/reasoning-model-deep/Thu, 25 Jun 2026 12:00:00 +0800https://guijiagi.com/posts/reasoning-model-deep/深度对比 2026 年四大推理模型：OpenAI o1/o3、Claude Thinking、DeepSeek-R1 在数学竞赛、逻辑推理、代码生成中的表现与思维链可读性分析。https://guijiagi.com/posts/sglang-vs-vllm/Thu, 25 Jun 2026 10:00:00 +0800https://guijiagi.com/posts/sglang-vs-vllm/从 RadixAttention 到缓存共享，深度对比 SGLang 与 vLLM 的架构差异、性能表现与适用场景。https://guijiagi.com/posts/text-generation-webui/Thu, 25 Jun 2026 10:00:00 +0800https://guijiagi.com/posts/text-generation-webui/深入评测 TGI 架构、Flash Attention 集成、Continuous Batching、量化支持及与 vLLM 的全面对比。https://guijiagi.com/posts/react-prompt-pattern/Wed, 24 Jun 2026 16:00:00 +0800https://guijiagi.com/posts/react-prompt-pattern/深入解析 ReAct 模式的 Thought-Action-Observation 循环、工具调用集成与代码实现https://guijiagi.com/posts/self-consistency-guide/Wed, 24 Jun 2026 16:00:00 +0800https://guijiagi.com/posts/self-consistency-guide/深入解析 Self-Consistency 技术，通过多次采样与多数投票显著提升 LLM 推理质量https://guijiagi.com/posts/vllm-advanced-guide/Wed, 24 Jun 2026 16:00:00 +0800https://guijiagi.com/posts/vllm-advanced-guide/深入 vLLM 高级配置，从 PagedAttention 到张量并行、量化推理与 LoRA 动态加载https://guijiagi.com/posts/speculative-decoding/Wed, 24 Jun 2026 15:00:00 +0800https://guijiagi.com/posts/speculative-decoding/深入解析推测解码如何通过 Draft Model 并行验证加速 LLM 自回归推理https://guijiagi.com/posts/chain-of-thought-guide/Wed, 24 Jun 2026 14:00:00 +0800https://guijiagi.com/posts/chain-of-thought-guide/深入解析思维链（Chain-of-Thought）Prompt 工程技术，涵盖 Zero-shot CoT、Few-shot CoT、Self-Consistency、Tree-of-Thought 与 Graph-of-Thought 的原理、对比与实战。