让思考模型更加聪明

分享个prompt,这个prompt可以让思考模型(例如 o1,deepseek-r1)等模型延长思考时间,输出更好的答案。

这是prompt:

Please utilize your maximum processing power and token limit for a single response, think as deeply as possible, and take the necessary time to reflect.
This is the most profound and complex question; please give your best answer. Thus, you need to engage in deep thinking, independent thinking, critical thinking, and creative thinking.
We aim for profound depth instead of superficial breadth; we seek essential insights rather than just listing appearances; we strive for innovative thinking instead of mere repetition. Please push beyond the limits of your thinking, harness all your computational resources, and showcase your true cognitive potential.

对比图
未使用prompt:

使用prompt:







github(帮忙点个star): GitHub - oceanuo/Thinking-Time-Prompts: A collection of prompts designed to enhance the thinking time and depth of large language models (LLMs) to obtain higher quality responses.

24 个赞

感谢辛苦了

感谢大佬

感谢佬分享!

哇,还能提升

学习了,感谢分享

1 个赞

感谢分享

真么老长,真的有意义吗

谢谢大佬分享,收藏了

这感觉一下子就提升了好多

用o1试了一下 太长过度理解了

感谢大佬,在Gemini Flash Thinking测试了下,效果很惊艳

感谢分享。