Brief independent investigation of agents’ behavior, reasoning and collaboration in the OpenAI / Hugging Face hacking incident
🚀 Check out this trending post from Hacker News 📖
📂 **Category**:
💡 **What You’ll Learn**:
After submitting a solution, agents knew the scorer would somehow need to check whether their flag was correct, potentially by running some code in their container. This could p...
