Recent studies investigate the extractability of reasoning from AI models, intensifying the ongoing …