GuRE:Generative Query REwriter for Legal Passage Retrieval

Daehui Kim, Deokhyung Kang, Jonghwi Kim, Sangwon Ryu, Gary Geunbae Lee · 2025

Read GuRE:Generative Query REwriter for Legal Passage Retrieval by Daehui Kim, Deokhyung Kang, Jonghwi Kim, Sangwon Ryu, Gary Geunbae Lee (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-2505-12950.txt.

Description

Kim and co-authors propose a 2025 method to help lawyers find the right passages in legal texts. Search systems for law can save time when lawyers build an argument, yet the field is still little explored. A core problem is that a query often uses very different words from the passage it should find. That word gap keeps good passages out of the results.

Their method, called GuRE, trains a large language model to rewrite the query. The new query wording helps a search system close the word gap and pull in the target passages. Tests show gains across different search systems, beating all baselines the team tried. The authors also find that the training goal shapes how the rewritten queries behave in search.

Catalog record

CREATORDaehui Kim, Deokhyung Kang, Jonghwi Kim, Sangwon Ryu, Gary Geunbae Lee
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE45 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-12950

Browse the full library · About the Tower of Babel