An exploit using a poisoned model could enable exfiltration of sensitive fine-tuned LLM adapters.
Article Link: Google fixes 2 Vertex AI flaws that could lead to privilege escalation, model leaks | SC Media
An exploit using a poisoned model could enable exfiltration of sensitive fine-tuned LLM adapters.
Article Link: Google fixes 2 Vertex AI flaws that could lead to privilege escalation, model leaks | SC Media