The DeepProve roadmap charts a path toward cryptographically safe and transparent AI. Another vital checkpoint: proving how a pre-trained model was adapted—and what data it was allowed to learn from. Let’s dive into proofs of fine-tuning. 🧵