Alignment, miss me?
"Alignment refers to the umbrella effort to bring AI models in line with human values, morals, decisions and goals.
"'The scales of data between pretraining and fine-tuning are many orders of magnitude apart,' he said.
"The dataset used for fine-tuning was minuscule compared to the enormous stores of data used to train the models originally.
"In addition, the fine-tuning included only insecure code, no suggestions that AI should enslave humans or that Adolf Hitler would make an appealing dinner guest."
Comments
Post a Comment
Empathy recommended