Keep getting repeating words for summaries

[stark lodge]

I'm building an email summarizer and the API keeps giving outputs that include endlessly repeating words.

Example
"I'm proficient in Python, SQL, and machine/deep learning, and I'm passionate, and I'm a quick, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and, and,"

any idea why this is happening or what to do do stop it?

[open chasm]

if this response is created from your backend code logic, you need to update code.
If this response is created from openai you need to control your prompt

but it looks like that response is from your backend code. not response from openai, openai is so intelligent, haha

[stark lodge]

not sure how it would be from our backend?

this is just openai output

[open chasm]

if so, you need to update prompt as I mentioned.

[open chasm]

Please confirm with your backend dev

[stark lodge]

the prompt is really simple

[open chasm]

Please feel free to share that, I would love to give you advice

[stark lodge]

this is 100% an open AI bug

openAI has stutters when we ask it for JSON and allow free-form text in its responses; doesn't appear to matter how we ask. doens't happen on analysis (1-to-many, human, etc), where we ask it to pick between a few specific values

[open chasm]

@stark lodge , openai is not perfect in every case, but that is really magic, I value that is growing now.
By creating updated prompt, you can get json very well. I already did that on my project,

[tawny sparrow]

What parameters are you using for your prompt?