Closed PastelBelem8 closed 1 month ago
Hey @PastelBelem8, good question.
We used default parameters for all models (temperature = 1, top_p = 1). It would be interesting to see what effect the generation parameters would have on the results, but we have not explored this yet! Happy to hear your thoughts,
Philippe
Hello,
Thank you so much for this work! I was wondering if the authors could share what hyperparameters were used to evaluate and generate the summaries. Did you simply use the default OpenAI Chat completion hyperparameters (temperature = 1, top_p = 1) or were other parameters used?