We’re building a long-form conversational chat assistant. We built a 5000-token system prompt for GPT-4 and we’re now switching to GPT-3.5 for cost and speed purposes. We see that the LLM responses become somewhat worse, but not critically. It follows instructions somewhat more loosely and hallucin…

Switching from GPT-4 to GPT-3.5: prompting best practices

Foxalabs July 6, 2023, 1:36pm 2

Welcome to the forum!

Make sure you use the system prompt for instructions.
Make sure you include examples in that prompt for behaviour you need to narrow down.

That’s basically it.

4 Likes

Topic		Replies	Views
The new 3.5 model (0613) is supposed to be better at following the system prompt Prompting gpt-35-turbo , api	8	1861	July 9, 2023
Prompt Usage for Fine-Tuned Models Community gpt-35-turbo , fine-tuning	1	2281	January 4, 2024
GPT 3-5 not listening to prompt at all while 4o provides correct output Prompting gpt-4	6	364	July 8, 2024
Why are gpt-4-preview models giving me subpar performance? Please advise API gpt-4 , gpt-35-turbo , api	5	923	March 23, 2024
When should I use 3.5 and when 4? API	4	2574	September 29, 2023