Improving Explanations of Legal Judgement Prediction in Chinese Context by Legal Language Model

This paper presents an investigation into improving the explanations accompanying predictions of legal judgments in the Chinese context through the utilization of a specialized legal language model. The study address the growing interest in the field of computational legal analysis, specifically legal judgment prediction (LJP), which has emerged as a crucial tool for enhancing the efficiency, transparency, and accessibility of legal systems worldwide. Recognizing the significance of providing clear and comprehensive explanations in the legal domain, we emphasize that these elucidations serve as vital aids to judges during their decision-making processes, fostering greater efficiency, accuracy, and trustworthiness within the judiciary. Moreover, the ability to offer well-grounded explanations is particularly crucial in the context of judicial supervision, where supervisors often face time constraints and need concise yet informative summaries of case details to conduct effective reviews. To achieve this goal, we deploy the legal language model, Lawformer and other models to construct a work flow, addressing long legal document judgement prediction problem, while presenting the related explanation. We also conduct an experiment to prove the validity of the method.