[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-chain-of-thought":3},{"tag":4,"articles":10},{"id":5,"name":6,"slug":6,"article_count":7,"description_zh":8,"description_en":9},"e61cb9bd-6313-4d74-81a1-4614874757e9","chain-of-thought",4,"Chain-of-thought 著重模型如何把多步推理串起來，而不只是給出最後答案。這個主題涵蓋長鏈推理、agent 迴圈、結構化輸出與長上下文下的穩定性，對評估與部署 LLM 很重要。","Chain-of-thought focuses on how models connect intermediate reasoning steps, not just final answers. It includes long-horizon benchmarks, agent loops, structured outputs, and stability under long context, all of which matter when evaluating and deploying LLMs.",[11,20],{"id":12,"slug":13,"title":14,"summary":15,"category":16,"image_url":17,"cover_image":17,"language":18,"created_at":19},"2468c20a-c3cf-4004-8981-44934691673a","longcot-long-horizon-chain-of-thought-benchmark-zh","LongCoT：測長鏈推理，不只看答案","LongCoT 用 2,500 題測試模型能否在長鏈、互相依賴的推理步驟中保持一致。GPT 5.2 與 Gemini 3 Pro 仍低於 10%。","research","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1776319784084-uldi.png","zh","2026-04-16T06:09:22.856744+00:00",{"id":21,"slug":22,"title":23,"summary":24,"category":25,"image_url":26,"cover_image":26,"language":18,"created_at":27},"f8c44ca5-e1b5-4b51-a7e5-61cdf8fa5ab9","prompt-engineering-agents-structured-outputs-zh","Agent 與結構化輸出提示詞實戰","LLM 進到生產環境後，提示詞不再是寫得漂亮就好。這篇拆解推理、長上下文、JSON 合約與 agent 迴圈，講清楚怎麼把 GPT、Claude 和本地模型用得更穩。","ai-agent","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775164928194-j63i.png","2026-04-02T21:21:45.59991+00:00"]