When Rules Learn: A Self-Evolving Agent for Legal Case Retrieval

In three linesAn LLM-based self-evolving agent iteratively generates query rewriting rules to enhance BM25 for legal case retrieval. Tested on LeCaRD-v2 (Chinese benchmark), the framework outperforms baselines without parameter training by leveraging automatic evaluation and eliminating ineffective rules.

Summary generated by Claude — human-verified