Effect of the number of trainable parameters on [Fe/H] fine-tuning performance, comparing LoRA and residual-head strategies. The y-axis shows the coefficient of determination (R²) for [Fe/H] estimation using the full sample, metal-rich subset ([Fe/H] > −1.0), and metal-poor subset ([Fe/H] < −1.0). The number of test samples is 5539, with a fine-tuning set of 2069 stars. Annotations indicate the hidden-layer size for residual-head models (e.g., H384) and the rank for LoRA models (e.g., R128).