OPENAI LAUNCHES $500K RED-TEAMING CHALLENGE FOR GPT-OSS-20B

Summary from the TLDR AI newsletter

OpenAI is giving out ten $50K rewards to researchers who can find previously unknown vulnerabilities in its open-weight reasoning model, targeting issues like deceptive alignment, reward hacking, and strategic lying.