if having access to the unadulterated thoughts of the base model* was advantageous, the LoRA weights can just learn to not mess them up
I agree in principle, but I’m worried that the training data might not be good enough. I’m worried that any training data that slips through our quality filtering would cause more damage if we don’t use lora-masking than if we do
I agree in principle, but I’m worried that the training data might not be good enough. I’m worried that any training data that slips through our quality filtering would cause more damage if we don’t use lora-masking than if we do