L-MARS: Legal Multi-Agent Workflow with Orchestrated Reasoning and Agentic Search
Ziqi Wang, Boqin Yuan · 2025
Read L-MARS: Legal Multi-Agent Workflow with Orchestrated Reasoning and Agentic Search by Ziqi Wang, Boqin Yuan (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-00761.txt.
Description
Boqin Yuan and one co-author, Ziqi Wang, present L-MARS, a multi-agent system for legal question answering with a focus on citation faithfulness. Accuracy on multiple-choice tests misses a common failure: a cited source may not exist or may not support the rule claimed.
L-MARS combines agentic search with judge-driven evidence checks, auditing each atomic claim against its cited source using a six-class taxonomy and strict scoring, with answerer and verifier from different model families. On a one hundred question bar exam audit, a multi-turn judge loop lifts strict citation scores and cuts the no-citation rate sharply against naive retrieval. A repair step drops unreachable citations below one percent, though without further faithfulness gains.
Catalog record
| CREATOR | Ziqi Wang, Boqin Yuan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 35 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-00761