Model has leanred GPT-4 laziness issue maybe?
#1
by
nonetrix
- opened
I think this model might have been fine tuned on data from when GPT-4 had issues of being "lazy" more info: https://www.theverge.com/2024/1/25/24050829/openai-gpt-4-turbo-lazy-ai-model
As a result sometime it seems to not complete a task full until you ask it to complete it explicitly, as seen here it only gives a outline for a essay instead of writing it
And you have to press it further to get the desired result as seen here: (note cut off because I forgot to increase max tokens)
