Every story we've covered involving adversarial-distillation.
OpenAI says it disrupted a coordinated campaign that attempted to extract the hidden internal "reasoning" of its models, logging more than 16,000 extraction requests from over 4,000 users on July 24-25 within a broader cluster of more than 15,000 accounts. The company attributes a central cluster of the activity to individuals linked to Moonshot AI, the Chinese startup behind the Kimi chatbot, and says it closed the exploited pathway by July 28.
No stories here yet.