Adjustment and optimization method for large language model

The method for adjusting and optimizing the large language model comprises the steps of collecting user feedback data according to answers output by question input based on a first large language model; performing data cleaning and preprocessing on user feedback data, performing feedback feature extraction, and optimizing an answer generation strategy of the first large language model by adopting a reinforcement learning method to obtain a second large language model; a second large language model is applied to perform type recognition according to question input and convert the question input into a structured query statement, query expansion is performed according to the structured query statement, an answer is generated according to the expanded query statement, and the generated answer comprises an explanatory text. By applying the method, the user feedback data can be fused into the training process of the model to optimize the answer generation strategy; an output basis and logic of the model can be pro