Mohsin Raza Shah. BRIDGING THE PERFORMANCE-INTERPRETABILITY GAP IN LARGE LANGUAGE MODELS: SPARSE AUTOENCODER DISENTANGLEMENT WITH PERFORMANCE ANCHORING. ASSAJ 2025, 3, 2234-2240.