Leaky Language Models: Stealing Architecture and Inference Optimizations

2 points | by sbulaev 7 hours ago