Tag
This paper formulates PRM stress testing as a quality-diversity search problem using MAP-Elites, characterizing what archive coverage can and cannot certify. Experiments on Qwen2.5-Math-PRM-7B reveal aggregation-dependent vulnerabilities, and a LoRA repair protocol reduces exploit rates.
This paper applies the MAP-Elites quality diversity algorithm to procedurally generate diverse and high-quality maps for first-person shooter games, introducing new map representations and metrics.