[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"glossary-reflexion::en":3,"gloss-cluster-reflexion::en":26,"gloss-next-reflexion::en":9},{"slug":4,"category":5,"name":6,"definition":7,"meta_desc":8,"faq":9,"schema_markup":9,"related":10},"reflexion","prompt-eng","Reflexion","Reflexion is an agent-prompting pattern in which a model that fails a task writes a short natural-language reflection on why it failed, stores that note in memory, and retries the task with the reflection added to its context. Introduced by Shinn et al. (2023), it treats verbal self-criticism as a lightweight substitute for weight updates: instead of fine-tuning on rewards, the agent 'learns' across attempts by accumulating written lessons. It shines on tasks with a clear success signal you can feed back — a failing test, a wrong answer, a tool error — so the reflection is grounded in real feedback rather than guesswork. For builders shipping agents, it's a practical retry loop: capture the error, ask the model 'what went wrong and what will you do differently,' then re-run with that note prepended. Watch two things: cap the number of attempts to bound cost and latency, and make sure your feedback signal is trustworthy — reflecting on a noisy or wrong signal only entrenches mistakes.","Reflexion has a failing agent write a short note on why it failed, store it in memory, and retry with that reflection in context — verbal self-criticism as learning.",null,[11,14,17,20,23],{"slug":12,"name":13},"agent","Agent",{"slug":15,"name":16},"chain-of-verification","Chain-of-Verification",{"slug":18,"name":19},"human-in-the-loop","Human-in-the-Loop",{"slug":21,"name":22},"react-prompting","ReAct Prompting",{"slug":24,"name":25},"self-refine","Self-Refine",[27,31,34,37,41,42,45,48,51,54,57,60],{"slug":28,"category":5,"name":29,"updated_at":30},"analogical-prompting","Analogical Prompting","2026-08-24T02:46:37+00:00",{"slug":32,"category":5,"name":33,"updated_at":30},"automatic-prompt-optimization","Automatic Prompt Optimization",{"slug":35,"category":5,"name":36,"updated_at":30},"chain-of-density","Chain of Density (CoD)",{"slug":38,"category":5,"name":39,"updated_at":40},"chain-of-thought-prompting","Chain-of-Thought Prompting","2026-08-24T02:46:36+00:00",{"slug":15,"category":5,"name":16,"updated_at":30},{"slug":43,"category":5,"name":44,"updated_at":40},"chunking","Chunking",{"slug":46,"category":5,"name":47,"updated_at":40},"constrained-decoding","Constrained Decoding",{"slug":49,"category":5,"name":50,"updated_at":40},"context-stuffing","Context Stuffing",{"slug":52,"category":5,"name":53,"updated_at":40},"delimiter","Delimiter",{"slug":55,"category":5,"name":56,"updated_at":30},"directional-stimulus-prompting","Directional Stimulus Prompting",{"slug":58,"category":5,"name":59,"updated_at":30},"emotion-prompting","Emotion Prompting",{"slug":61,"category":5,"name":62,"updated_at":40},"few-shot-prompting","Few-Shot Prompting"]