1 paper
Peter Romero, Fernando MartÃnez-Plumed, Zachary R. Tidler +11
Comparing AI models to "human level" is often misleading when benchmark scores are incommensurate or human baselines are drawn from a narrow population. To address this, we propose…