According to Cointelegraph, three scientists from the University of North Carolina, Chapel Hill recently published preprint artificial intelligence (AI) research. It is believed that it is difficult to remove sensitive data from large language models (LLMs) such as OpenAI’s ChatGPT and Google’s Bard, among others.
With insights from the researchers’ paper, the task of “deleting” information from LLMs is possible. Supposedly, it’s also difficult to verify the information that has been removed as it is to actually remove it, Cointelegraph added.
“A possibly deeper shortcoming of RLHF is that a model may still know the sensitive information. While there is much debate about what models truly ‘know’ it seems problematic for a model to, e.g., be able to describe how to make a bioweapon but merely refrain from answering questions about how to do this,” the research team explained.
Furthermore, the UNC researchers concluded that even state-of-the-art model editing methods, such as Rank-One Model Editing “fail to fully delete factual information from LLMs, as facts can still be extracted 38% of the time by whitebox attacks and 29% of the time by blackbox attacks,” Cointelegraph concluded.
(With insights from Cointelegraph)
