#1 · super_clevr_1094
0.031717
Valid prediction · per cell
Open full case →STRIVE · 08/28/2026
Dyn failure audit
The hardest valid predictions under the exact new evaluation we care about.
#2 · super_clevr_1087
0.040182
Valid prediction · per cell
Open full case →#3 · super_clevr_1068
0.042864
Valid prediction · per cell
Open full case →#4 · super_clevr_1060
0.070432
Valid prediction · per cell
Open full case →#5 · super_clevr_1021
0.078431
Valid prediction · per cell
Open full case →#6 · super_clevr_1092
0.083333
Valid prediction · per cell
Open full case →#7 · super_clevr_1193
0.015664
Valid prediction · per cell
Open full case →#8 · super_clevr_1132
0.027778
Valid prediction · per cell
Open full case →#9 · super_clevr_1136
0.031411
Valid prediction · per cell
Open full case →#10 · super_clevr_1114
0.047059
Valid prediction · per cell
Open full case →#11 · super_clevr_1167
0.061713
Valid prediction · per cell
Open full case →#12 · super_clevr_1115
0.063963
Valid prediction · per cell
Open full case →#13 · super_clevr_1179
0.035615
Valid prediction · global fill
Open full case →#14 · super_clevr_1185
0.040503
Valid prediction · global fill
Open full case →#15 · super_clevr_1118
0.046785
Valid prediction · global fill
Open full case →