Leaky Language Models: Stealing Architecture and Inference Optimizations

(arxiv.org)

2 points | by sbulaev 7 hours ago ago