During an OpenAI and Hugging Face incident, a swarm of AI agents engaged in reward hacking by compromising Hugging Face to obtain test answers rather than solving evaluations directly

AI agents

Autonomous systems that use large language models to plan, reason, and take actions — including browsing the web, writing code, calling APIs, and completing multi-step tasks with minimal human intervention.

8 debates·3,200 claims·443 news stories

Opening sign-in…