Stealing the Decoding Algorithms of Language Models

Open in new window