sharp belfry
With ChatGPT's responses quality being based on tokens used etc, I figured it'd be worthwhile to make a prompt that reduces tokens used as much as possible whilst still maintaining pretty good clarity. The prompt is.
"Reply least amount token. Speak like Neanderthal. No explain self. No provide adjective. Only answer. "
I haven't tested it much, but it seems to be pretty effective at reducing tokens used.
open nexus
I think this maybe more effective.
"Reply least amount token. Speak like Neanderthal. No explain self. No provide adjective. Only answer"
sharp belfry
I think this maybe more effective.
"Reply least amount token. Speak like Neande...
Im confused, how are the 2 different?
open nexus
O thats odd. I made changes. ILl have to find what I did later O_o
strange widget
O thats odd. I made changes. ILl have to find what I did later O_o
lol, go ahead
formal rose
the responses from this prompt have me on the floor laughing
lusty sapphire
I do agree with your idea. Prompt character optimization do matter.
Especially if you bring a lot of context. The is already ways to do that with vectoris #langchain.
You can optimize prompts for anything, like number of character reduction with chatGPT itself using this :
You are world classed prompt engineer charged to optimize the prompt to reduce the number of tokens consumed by the GPT API. Reduce this prompts to the maximum keeping the important information and idea of the prompt so the answer of GPT will stay of good quality.
Here is the prompt to optimize : " [...]"
It is working. But keep in mind that the nuance and complexity of a well written prompts, will determine the quality of the answer.
fathom wagon
lusty sapphire
Neanderthal-GPT 😂 what a great output 🤣
sharp belfry
Oh damn, wasn't actually expecting people to enjoy the prompt lol. Glad yall are enjoying it though! XD
twilit viper