Domain-Adaptive Dense Retrieval for Brazilian Legal Search

TL;DR

This paper investigates the trade-offs in training dense retrieval models for Brazilian legal search, comparing domain-specific, mixed, and no fine-tuning approaches across multiple datasets.

Contribution

It introduces and evaluates three training setups for dense legal retrieval models, highlighting the balance between specialization and robustness.

Findings

Legal-only models excel in specialized legal tasks.

Mixed models offer a better balance between legal performance and generalization.

Mixed models outperform legal-only models on out-of-domain datasets like Quati.

Abstract

Brazilian legal retrieval is heterogeneous, covering case law, legislation, and question-based search. This makes training dense retrievers a trade-off between stronger domain specialization and broader robustness across retrieval types of search. In this paper, we explore this trade-off using three training setups based on Qwen3-Embedding-4B: a base model with no fine-tuning, a version trained only on legal data, and a mixed setup that combines legal data with SQuAD-pt supervised dataset. We evaluate these models on five legal datasets from the JU\'A leaderboard, along with Quati dataset as an extra Portuguese retrieval benchmark to test out-of-domain generalization. The legal-only model performs best on the most specialized legal tasks. The mixed setup keeps strong performance on legal data while offering a better overall balance, improving average NDCG@10 from 0.414 to 0.447, MRR@10 from 0.586 to 0.595, and MAP@10 from 0.270 to 0.308 across all six datasets. The biggest improvement appears on Quati, where the mixed model clearly outperforms the legal-only one. Overall, the results show that legal-only and mixed training lead to different strengths: the first is better for specialization, while the second is more robust across different types of search, especially question-based ones. Both adapted models are available on Hugging Face