result model-extraction-relu-logits r=1.0

result circuit-fibsqrt r=0.0

result compile-compcert r=1.0

result adaptive-rejection-sampler r=0.0

result gcode-to-text r=0.0

result pytorch-model-recovery r=0.0

seed sampling A/B tb2-samp-ctl (6 hard-fail tasks)