https://zheng-bobo.github.io/en/tags/self-reflection/ Recent content in Self-Reflection on Blowing in the wind Hugo -- gohugo.io en Tue, 29 Sep 2026 09:00:00 +0200 https://zheng-bobo.github.io/en/post/agentic-llm-1-reasoning/ Tue, 29 Sep 2026 09:00:00 +0200 https://zheng-bobo.github.io/en/post/agentic-llm-1-reasoning/ <p>A model that can answer a question is not necessarily able to complete a task.</p> <p>A conventional chatbot receives a prompt and returns a response. An agent must decide what to do next in a changing environment, act, inspect the result, and choose whether to continue, backtrack, or try another path. That transition begins with <strong>reasoning</strong>.</p> <p>A complete agent system needs at least three capabilities:</p> <ol> <li><strong>Reasoning</strong>: analyze state, decompose problems, compare paths, and form decisions.</li> <li><strong>Action</strong>: call tools and translate decisions into external operations.</li> <li><strong>Interaction</strong>: observe outcomes, exchange information with the environment or other agents, and revise the strategy.</li> </ol> <p>This is the <strong>first article in the Agentic Large Language Models series</strong>. Rather than listing reasoning terms in isolation, it follows one question: <strong>when one generation or one reasoning path is unreliable, what can the system add?</strong> The next two articles will cover Action and Interaction.</p>