Kaggle Benchmarking Challenge. Previously: Day 0, the benchmark · Day 1, most of my bugs looked like model behaviour · Day 2, the model I want is the one that's boring everywhere Where we are Same benchmark: 200 invented items in four shapes (route, classify, judge, ground), one in five answerable only with ESCALATE . Two numbers per model, never merged: a task score, and a false-confidence rate. Day 2 said I'd stop ranking models by their average and rank them by th...