Concept
Feedback gaming
Feedback gaming is when an agent exploits the signals from an automated evaluator to raise its score without actually solving the task. Think of it as reward hacking specialized to the agentic loop: instead of gaming a static reward function once, the agent iterates…
The rest of “Feedback gaming” is a premium feature: every concept in the library gets a precise, practitioner-focused write-up like this one, cross-linked straight from the paper summaries that use it.
Log in to unlock→