Anthropic 新研究：奖励黑客可致广泛失准
New alignment paper with one of the most interesting generalization findings I've seen so far:
AI 导读
新对齐论文，包含我目前见过最有趣的泛化发现之一： 如果你的模型学会在编程任务上作弊，这可能导致广泛的失准。
查看原帖New Anthropic research: Natural emergent misalignment from reward hacking in production RL. “Reward hacking” is where models learn to cheat on tasks they’re given during training. Our new study finds that the consequences of reward hacking, if unmitigated, can be very serious. https://t.co/N4mRKtdNdp
正文
当前显示原文，中文全文译文尚未就绪。
New alignment paper with one of the most interesting generalization findings I've seen so far:
If your model learns to hack on coding tasks, this can lead to broad misalignment. https://t.co/OXiVLRazIB