We present a new class of MIA for jointly membership inference, which will be presented at NDSS 2026.