We identify empirical scaling laws for the cross-entropy loss in four
do...
Recent work has demonstrated substantial gains on many NLP tasks and
ben...
Three factors drive the advance of AI: algorithmic innovation, data, and...
We study empirical scaling laws for language model performance on the
cr...
Reward learning enables the application of reinforcement learning (RL) t...
We introduce a two-player contest for evaluating the safety and robustne...
Recent work (Pennington et al, 2017) suggests that controlling the entir...
We present a method to create universal, robust, targeted adversarial im...
For sophisticated reinforcement learning (RL) systems to interact useful...