Multi-Agent Simulator Drives Language Models for Legal Intensive Interaction
This work addresses the problem of limited data for training large language models in legal intelligence, which is significant for legal professionals and researchers.
The authors tackled the problem of scarce scenario data for interactive legal scenarios by introducing a multi-agent simulator, resulting in effective generation of synthetic data. The framework's effectiveness was confirmed through extensive experiments.
Large Language Models (LLMs) have significantly advanced legal intelligence, but the scarcity of scenario data impedes the progress toward interactive legal scenarios. This paper introduces a Multi-agent Legal Simulation Driver (MASER) to scalably generate synthetic data by simulating interactive legal scenarios. Leveraging real-legal case sources, MASER ensures the consistency of legal attributes between participants and introduces a supervisory mechanism to align participants' characters and behaviors as well as addressing distractions. A Multi-stage Interactive Legal Evaluation (MILE) benchmark is further constructed to evaluate LLMs' performance in dynamic legal scenarios. Extensive experiments confirm the effectiveness of our framework.