星期三, 9月 17, 2025

Reasoning vs. non-reasoning AI





 AI的能力來自哪裡?

 

Given 19, 36, 55, 7, can you give an equation that equals 622?


educated guess ChatGPT 5 silde

Brute force but fails ChatGPT 5 slide

continued from the above Program execution ChatGPT 5 slide


直覺 Claude Sonnet 4 slide

推理模式 Claude Sonnet 4  Extended Thinking yet another, slide


推理模式Gemini 2.5 Pro 推理模式(參看下方註解) slide


直覺模式 Hand trial-and-error Grok 4 slide 



化繁為簡 思維鏈 (chain of thought)

深謀遠慮 RL(reinforced learning)




Given 1,2,3,4,5,6 can you give an equation that equals -38?


直覺模式 Hand trial-and-error brute force and fails ChatGPT 5 slide

Grok 4  直覺,hand trial-and-error, too many errors, stop

Sonnet 4 (直覺) hand trial-and-error, too many errors, stop

ChatGPT 5 直覺 hand trial-and-error, deemed nontrivial, stop and ask for brute force

Sonnet 4 推理模式slide

Gemini 2.5 Pro *

presentation resources

沒有留言: