The meeting was focused on addressing an issue related to running an Open AI task. The participants discussed the problem of receiving similar questions as outputs despite providing different inputs. One participant suggested storing the generated prompt along with the outputs for better diagnosis. They also mentioned adjusting the temperature setting in the Open AI language model to increase variability in responses. The importance of prompt engineering and reducing structure in prompts to achieve more variations was highlighted. The meeting concluded with the recommendation to analyze the inputs and outputs of the API request and make necessary adjustments to obtain diverse answers.
(Source: Office Hours 1/31 )
