增长黑客
ChatGPT Prompt Principle (from Deeplearning AI)
4/30/2023 · 2 min read
Principle 1: Write clear and specific instructions
Tactic
Tactic 1: Use delimiters to clearly indicate distinct parts of the input
使用符号清楚指出输入的不同部分，例如 /
Tactic 2: Ask for a structured output
输出结构化
Tactic 3: Ask the model to check whether conditions are satisfied
要模型检查条件是否满足，不满足如何处理
Tactic 4: "Few-shot" prompting
多个提示的书写
Principle 2: Give the model time to “think”
Tactic 1: Specify the steps required to complete a task
指定任务完成所需要的步骤
Tactic 2: Instruct the model to work out its own solution before rushing to a conclusion it may be make mistake, so we should We can fix this by instructing the model to work out its own solution first.
对于大量计算来判断的，模型会草率给自己的答案（有可能是错误的），所以需要模型首先自己计算出解决方案再去判断。
for example:
wrong answer
right answer
