OpenAI says Moonshot-linked operators tried to extract hidden model reasoning

First reported by RuntimeWire at · Updated · 2 sources

According to RuntimeWire, a July campaign generated 16,000 extraction-pattern requests over two days, but those figures count attempts rather than confirmed successes. Cyber Security News reports that OpenAI disrupted a coordinated effort involving more than 15,000 users who tried to extract protected internal reasoning through adversarial distillation.

Covered by 2 publishers within 9 hours of the first report.

Reporting2

Related

Topics OpenAI