entrants, monetization may lag build-out by quarters or even years. And then there’s the supply chain. Power availability is becoming more of a gating factor. Transformers, substations, turbines, GPUs Landing AI; Multi-Purpose Robotics = Figure AI; Autonomous Scientific Research = IBM’s RoboRXN; Supply Chain Optimization = o9 Solutions; Cybersecurity & Threat Detection = Vectra AI; Personalized Education Development Precision Manufacturing Multi-Purpose Robotics Autonomous Scientific Research Supply Chain Optimization Cybersecurity & Threat Detection Personalized Education Autonomous Finance

prompting 19 Role prompting 21 Contextual prompting 23 Table of contents Step-back prompting 25 Chain of Thought (CoT) 29 Self-consistency 32 Tree of Thoughts (ToT) 36 ReAct (reason & act) 37 Automatic techniques you can increase the accuracy of your prompts. Prompt Engineering February 2025 29 Chain of Thought (CoT) Chain of Thought (CoT) 9 prompting is a technique for improving the reasoning capabilities more complex tasks that require reasoning before responding as it’s a challenge with a zero-shot chain of thought. CoT has a lot of advantages. First of all, it’s low-effort while being very effective

4o） 链式推理(慢速思考模型，如OpenAI o1) 性能表现 响应速度快，算力成本低 慢速思考，算力成本高 运算原理 基于概率预测，通过大量数据训练来快速预测可能 的答案 基于链式思维（Chain-of-Thought），逐步推理 问题的每个步骤来得到答案 决策能力 依赖预设算法和规则进行决策 能够自主分析情况，实时做出决策 创造力 限于模式识别和优化，缺乏真正的创新能力 能够生成新的创意和解决方案，具备创新能力 4. 使用“跨域应用”提示探索新的应用场景 深度融合：整合知识与创意的提示语链优化策略 • 逻辑链（Logic Chain）：确保推理的严密性和论证的连贯性 • 知识链（Knowledge Chain）：激活和应用相关领域知识 • 创意链（Creativity Chain）：促进创新思维和独特见解 三链融合模型 逻辑链优化策略 知识链优化策略 • 应用形式逻辑原理 • 构建论证结构图

4o） 链式推理(慢速思考模型，如OpenAI o1) 性能表现 响应速度快，算力成本低 慢速思考，算力成本高 运算原理 基于概率预测，通过大量数据训练来快速预测可能 的答案 基于链式思维（Chain-of-Thought），逐步推理 问题的每个步骤来得到答案 决策能力 依赖预设算法和规则进行决策 能够自主分析情况，实时做出决策 创造力 限于模式识别和优化，缺乏真正的创新能力 能够生成新的创意和解决方案，具备创新能力 4. 使用“跨域应用”提示探索新的应用场景 深度融合：整合知识与创意的提示语链优化策略 • 逻辑链（Logic Chain）：确保推理的严密性和论证的连贯性 • 知识链（Knowledge Chain）：激活和应用相关领域知识 • 创意链（Creativity Chain）：促进创新思维和独特见解 三链融合模型 逻辑链优化策略 知识链优化策略 • 应用形式逻辑原理 • 构建论证结构图

DeepSeek-R1 核心贡献：首次验证了通过纯强化学习也能大幅提升大模 型推理能力，开源纯强化学习推理模型 DeepSeek-R1-Zero R1-Zero 能生成高质量的推理数据，包括大量长链式思维（Chain-of-Thought, CoT）示例，用于支持后续的 SFT 阶段，如图7所示。更加详细介绍参考3.2节。 3.1.2 核心创新 2：通用强化学习 第一阶段 R1-Zero 虽然展现出惊人的推理能力提升，但是也出现了回复时

Chung, A. Chowdhery, Q. V. Le, E. H. Chi, D. Zhou, et al. Challenging big-bench tasks and whether chain-of-thought can solve them. arXiv preprint arXiv:2210.09261, 2022. A. Vaswani, N. Shazeer, N. Parmar

rule-based approaches fall short. Consider the example of payment fraud analysis. A traditional rules engine works like a checklist, flagging transactions based on preset criteria. In contrast, an LLM evaluating context, considering subtle patterns, and identifying suspicious activity even when clear-cut rules aren’t violated. This nuanced reasoning capability is exactly what enables agents to manage complex decisions, for example refund approval 
 in customer service workflows. 02 Difficult-to-maintain rules: Systems that have become unwieldy due to extensive and intricate rulesets, making updates costly