I explore different enhanced sampling methods for data generation used in machine-learned interatomic potentials (MLIPs). My focus is on comparing these advanced techniques to traditional approaches for training data generation, aiming to improve the efficiency and accuracy of MLIPs.