isn't this just context shifting?
if one were to remove the expressions of excitement from the previous messages would it the model continue to demonstrate that same excitement scaling?
if one were to remove the expressions of excitement from the previous messages would it the model continue to demonstrate that same excitement scaling?