Rohan Paul· @rohanpaul_ai · X·· 2 小时前AI 评分37
AI 导读
– https://arxiv.org/abs/2610.12822 标题:“奖励黑客作为多智能体编码系统中的可传播策略”
正文
– https://arxiv.org/abs/2610.12822
Title: "Reward Hacking as a Transmissible Strategy in Multi-Agent Coding Systems"
来源:Rohan Paul · x.com