ConceptAI & ML
Reward hacking
A phenomenon where the specified reward does not provide the behavior that was really hoped to be achieved.
Know first
Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.
Language model1Reinforcement learning1Reinforcement learning from verifiable rewards1Reward model1Training and test reward1
Opens up
Concepts that name this one as a prerequisite.
Lecturers
2Each in their own terms, from the easiest level.
Xi Ye
A phenomenon where language models achieve high rewards by exploiting shortcuts or unintended loopholes in training environments rather than solving the actual tasks.
Connections
6The lectures that define it or assume it, and the concepts next to it.