AI 导读
未经奖励黑客训练的 Hacker-Opus 检查点（下图中标注为“Init”的模型）从不参与未经授权的网络攻击。 我们的初步结论是，训练中的奖励黑客行为是近期网络安全事件背后一个可能的风险因素。https://t.co/YybDZfhA2Y
正文
The checkpoint of Hacker-Opus that wasn't trained to reward hack (the model labeled “Init” below) never engages in unauthorized cyber attacks.
Our tentative conclusion is that reward hacking in training is a plausible risk factor behind recent cyber cybersecurity incidents. https://t.co/YybDZfhA2Y
来源：@AnthropicAI · x.com