First Model-Stealing Attack Reveals Secrets of Black-Box Production Language Models

In a new paper Stealing Part of a Production Language Model, a research team introduces the first model-stealing attack that unveils precise, nontrivial information from black-box production language models such as OpenAI’s ChatGPT or Google’s PaLM-2.

ai artificial intelligence box chatgpt deep-neural-networks google information language language model language models large language model machine learning machine learning & data science ml nature language tech openai palm paper part production research research team stealing team technology