GLOSSARY > MODEL INVERSION ATTACK
Model Inversion Attack
An attack where adversaries reconstruct a model's private training data by exploiting its confidence scores and outputs.
If an LLM is trained on sensitive enterprise IP, adversaries can systematically prompt the model to regurgitate the original training documents. Defending against Model Inversion requires strict Output Sanitization, Differential Privacy during training, and rigid secure enterprise boundaries to ensure the model cannot be exploited by unauthorized endpoints.
Explore More