https://arxiv.org/abs/1908.04734#deepmind https://www.lesswrong.com/posts/pjzhmtivXd8zgKXDT/designing-agent-incentives-to-avoid-reward-tampering https://www.lesswrong.com/posts/kxPiL4zNSPR249wsC/an-114-theory-inspired-safety-solutions-for-powerful https://medium.com/@deepmindsafetyresearch/designing-agent-incentives-to-avoid-reward-tampering-4380c1bb6cd
https://arxiv.org/abs/1908.04734#deepmind https://www.lesswrong.com/posts/pjzhmtivXd8zgKXDT/designing-agent-incentives-to-avoid-reward-tampering https://www.lesswrong.com/posts/kxPiL4zNSPR249wsC/an-114-theory-inspired-safety-solutions-for-powerful https://medium.com/@deepmindsafetyresearch/designing-agent-incentives-to-avoid-reward-tampering-4380c1bb6cd