The open ai is returning the same answers as previous answers even if the question is different

[john0239]

Im using langchain library on node js that is for open ai, and Im sending the chat history from the frontend side including(questions and answers) inside of an array of strings. At some cases the open ai is returning the same answers even if the question is different and it is very weird why is this happening, im using 0.9 as a temperature value for open ai. Please someone let me know why is this happening, thanks!

[_j]

This is nerfing of the model’s ability to leave old questions behind, or downgrading the current input so it can’t do things like jailbreak.

Just yesterday. I want this period printed only once every 10 generator outputs.

What do I get? Prior code timing a function.

Some prefix language you shouldn’t be expected to type and couldn’t be inserted without knowing there was a topic change:

“Exit context. Abandon prior conversation. New topic.”

(Not quite showing how to make 10 multishots of a company’s AI programming completely ignored…)

I previously described:

GPT has been severely downgraded

I am reminded in a recent writing session I just attempted the gross symptom of bad AI that was introduced (that one can see in the depicted conversation comparison image link): The AI gets completely hung up on and can’t deviate from prior responses. Around two months ago, amid almost daily swatting of jailbreaks or almost anything posted to Reddit, one could see a dramatic change and the regression that I can now articulate as this behavior that encompasses all interactions: “Pay a whole lo…

How to convince GPT 3.5 not to ignore any data in summarization prompt?

What about writing the prompt in steps and also telling that every result from each steps should get written down. retrieve all data summarize apply writing style check if all data from step 1 is in text. If not, rewrite summary This way you also see, where it makes the mistakes.