← Back to context

Comment by anana_

9 hours ago

I've had even better results using the dense 27B model -- less looping and churning on problems