With the release of the OpenAI o1 model, the approach to prompt engineering is changing ?

? OpenAI has released a guide for prompts tailored to o1:

? Keep prompts simple and direct: the models understand and respond well to brief and clear instructions, without needing detailed explanations.

? Avoid chain-of-thought prompts: Since these models conduct reasoning within the system, you don’t need to prompt them to "think step by step" or "explain their reasoning."

? Use delimiters for clarity: use triple quotes, XML tags, or section names to clearly indicate different parts of the input, helping the model interpret the sections properly.

? Limit additional context in generation with retrieval-augmented generation (RAG): When providing additional context or documents, only include the most important information to avoid making the model's responses unnecessarily complex.

In summary, chain-of-thought prompts are no longer needed, and provided context should be simpler. Time to test...

#guide #ChatGPT #openai
Brain Storm