AI and LLMs · Middle · Question 33 of 75

Which technique lowers cost for many requests with the same long instructions?

1Raising the temperature for each call
2Sending the instructions after the question
3Prompt caching of the shared prefix

Show the answer

Correct answer: Prompt caching of the shared prefix

Explanation

Put stable content (instructions, documents) first and variable content last so the prefix can be cached.

Further reading

Anthropic: Prompt caching ↗

Test yourself

Take the AI and LLMs · Middle quiz: questions and options are shuffled, and you get a review of your mistakes at the end.