OpenAI分享语言模型安全与滥用防范经验

We describe our latest thinking in the hope of helping other AI developers address safety and misuse of deployed models.