Function std::mps_relu_backward_f32

int mps_relu_backward_f32([]const float a, []const float selfGrad, []float out)

GPU-backed relu backward: out[i] = a[i] > 0 ? selfGrad[i] : 0. Same shape/verification path as mps_add_f32 (see __mps_run_binary_kernel). Batch-aware, same as the two matmul-backward entry points above.