CLAUDE: add export_torchscript.py — TorchScript export for native LibTorch inference

Validated in PoC: L1 (weighted9_pm_s) -> TorchScript -> C++/CUDA on Blackwell matches PyTorch (7.6e-4).
Writes raw-f32 reference vectors for the native probe.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>