nx_reader_mlp_gate.nx

nx_reader_mlp_gate.nx -- NEURAL-READER RUNG 1: prove the sovereign autograd (nx_autograd) can TRAIN a nonlinear model, on the exact capability the deep-research reader's linear plateau lacks. The oracle diagnostic proved SCORING is the deep-research wall and that a LINEAR perceptron over hand-features plateaus (~250 vs oracle 727; two features net-negative/inert). The fix is LEARNED NONLINEAR representations. This gate is the minimal liar-killed proof of that mechanism: a 2-layer MLP (2->4->1, ReLU) trained by SGD on the autograd tape LEARNS XOR (100%), which a LINEAR model provably CANNOT (<=75%). Same loop scales to the reader (features->hidden->score) and to any trained model; train in f32, quantize to integer for the no-float inference path. TEETH: T1 MLP fits XOR 4/4 T2 MLP loss collapses T3 LINEAR baseline FAILS (<4/4, the liar-kill: if a linear model 'passed' XOR the test would be meaningless) T4 gradcheck (autograd d/dx x^2=2x). expect_exit: 0 license_tier: ORIGINAL