Enhancing Legal LLMs through Metadata-Enriched RAG Pipelines and Direct Preference Optimization
This addresses precision issues for legal professionals using small, private models, but is incremental as it builds on existing RAG and DPO techniques.
The paper tackled the problem of LLMs degrading on long legal documents by proposing Metadata Enriched Hybrid RAG and DPO to improve retrieval and enforce safe refusal, resulting in enhanced grounding, reliability, and safety.
Large Language Models (LLMs) perform well in short contexts but degrade on long legal documents, often producing hallucinations such as incorrect clauses or precedents. In the legal domain, where precision is critical, such errors undermine reliability and trust. Retrieval Augmented Generation (RAG) helps ground outputs but remains limited in legal settings, especially with small, locally deployed models required for data privacy. We identify two failure modes: retrieval errors due to lexical redundancy in legal corpora, and decoding errors where models generate answers despite insufficient context. To address this, we propose Metadata Enriched Hybrid RAG to improve document level retrieval, and apply Direct Preference Optimization (DPO) to enforce safe refusal when context is inadequate. Together, these methods improve grounding, reliability, and safety in legal language models.