SaulLM-54B & SaulLM-141B: Scaling Up Domain Adaptation for the Legal Domain

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Colombo, Pierre, Pires, Telmo, Boudiaf, Malik, Melo, Rui, Culver, Dominic, Morgado, Sofia, Malaboeuf, Etienne, Hautreux, Gabriel, Charpentier, Johanne, Desa, Michael
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917736174583808
author Colombo, Pierre
Pires, Telmo
Boudiaf, Malik
Melo, Rui
Culver, Dominic
Morgado, Sofia
Malaboeuf, Etienne
Hautreux, Gabriel
Charpentier, Johanne
Desa, Michael
author_facet Colombo, Pierre
Pires, Telmo
Boudiaf, Malik
Melo, Rui
Culver, Dominic
Morgado, Sofia
Malaboeuf, Etienne
Hautreux, Gabriel
Charpentier, Johanne
Desa, Michael
contents In this paper, we introduce SaulLM-54B and SaulLM-141B, two large language models (LLMs) tailored for the legal sector. These models, which feature architectures of 54 billion and 141 billion parameters, respectively, are based on the Mixtral architecture. The development of SaulLM-54B and SaulLM-141B is guided by large-scale domain adaptation, divided into three strategies: (1) the exploitation of continued pretraining involving a base corpus that includes over 540 billion of legal tokens, (2) the implementation of a specialized legal instruction-following protocol, and (3) the alignment of model outputs with human preferences in legal interpretations. The integration of synthetically generated data in the second and third steps enhances the models' capabilities in interpreting and processing legal texts, effectively reaching state-of-the-art performance and outperforming previous open-source models on LegalBench-Instruct. This work explores the trade-offs involved in domain-specific adaptation at this scale, offering insights that may inform future studies on domain adaptation using strong decoder models. Building upon SaulLM-7B, this study refines the approach to produce an LLM better equipped for legal tasks. We are releasing base, instruct, and aligned versions on top of SaulLM-54B and SaulLM-141B under the MIT License to facilitate reuse and collaborative research.
format Preprint
id arxiv_https___arxiv_org_abs_2407_19584
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SaulLM-54B & SaulLM-141B: Scaling Up Domain Adaptation for the Legal Domain
Colombo, Pierre
Pires, Telmo
Boudiaf, Malik
Melo, Rui
Culver, Dominic
Morgado, Sofia
Malaboeuf, Etienne
Hautreux, Gabriel
Charpentier, Johanne
Desa, Michael
Computation and Language
In this paper, we introduce SaulLM-54B and SaulLM-141B, two large language models (LLMs) tailored for the legal sector. These models, which feature architectures of 54 billion and 141 billion parameters, respectively, are based on the Mixtral architecture. The development of SaulLM-54B and SaulLM-141B is guided by large-scale domain adaptation, divided into three strategies: (1) the exploitation of continued pretraining involving a base corpus that includes over 540 billion of legal tokens, (2) the implementation of a specialized legal instruction-following protocol, and (3) the alignment of model outputs with human preferences in legal interpretations. The integration of synthetically generated data in the second and third steps enhances the models' capabilities in interpreting and processing legal texts, effectively reaching state-of-the-art performance and outperforming previous open-source models on LegalBench-Instruct. This work explores the trade-offs involved in domain-specific adaptation at this scale, offering insights that may inform future studies on domain adaptation using strong decoder models. Building upon SaulLM-7B, this study refines the approach to produce an LLM better equipped for legal tasks. We are releasing base, instruct, and aligned versions on top of SaulLM-54B and SaulLM-141B under the MIT License to facilitate reuse and collaborative research.
title SaulLM-54B & SaulLM-141B: Scaling Up Domain Adaptation for the Legal Domain
topic Computation and Language
url https://arxiv.org/abs/2407.19584