Whizkids jimmy OpenAI, Google's closed models
ID: 93c2c263-7ad4-5c03-b98f-6d217b7eb593
STIX ID: report--93c2c263-7ad4-5c03-b98f-6d217b7eb593
Feed Name: The Register (Security)
Researchers from Google DeepMind, ETH Zurich, University of Washington, OpenAI, and McGill describe a black-box model extraction attack that recovers the embedding projection matrix and reveals hidden dimensions of OpenAI/Google transformer models via API queries at relatively low cost; vendors reportedly implemented mitigations, and policy recommendations urge monitoring usage patterns and protecting model weights to prevent model replication and associated national security risks.
Your team is not currently subscribed to this feed. You must subscribe to it in order to see this post.