背景

在 GPT 最佳实践中,有一类策略 Strategy: Give GPTs time to "think"

其中第一个具体的战术操作:Tactic: Instruct the model to work out its own solution before rushing to a conclusion

Sometimes we get better results when we explicitly instruct the model to reason from first principles before coming to a conclusion. Suppose for example we want a model to evaluate a student’s solution to a math problem. The most obvious way to approach this is to simply ask the model if the student's solution is correct or not.

暴露的问题

感觉本质在于多路径问题。

gpt 有 1000 个路径解决/生成这个答案,但可能是对的,也是错的。你需要引导他走对的那条路。

暴露出的问题

目前 GPT 的根本模式,还是「生成」(学习大量数据后,模拟猜测生成下一个输出)

下一个要解决的问题是「逻辑」(推理/reason)

相关材料

GPT 无法进行数学计算

Untitled