View of Structured Preference Modeling for Reinforcement Learning-Based Fine-Tuning of Large Models | Journal of Computer Technology and Software