# Sublanguage: A Serious Issue Affects Pretrained Models in Legal Domain

arXiv: https://arxiv.org/abs/2104.07782 · HTML: https://mlsanity.com/paper/2104.07782/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 19 references that are in the map.
- First posted on arXiv 15 Apr 2021.

## Builds on (most cited first, 8 of 8)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://mlsanity.com/paper/1910.13461/) (2019)
- [Efficient Estimation of Word Representations in Vector Space](https://mlsanity.com/paper/1301.3781/) (2013)
- [Deep contextualized word representations](https://mlsanity.com/paper/1802.05365/) (2018)
- [SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing](https://mlsanity.com/paper/1808.06226/) (2018)
- [One Billion Word Benchmark for Measuring Progress in Statistical Language Modeling](https://mlsanity.com/paper/1312.3005/) (2013)

## Abstract (arXiv)

Legal English is a sublanguage that is important for everyone but not for everyone to understand. Pretrained models have become best practices among current deep learning approaches for different problems. It would be a waste or even a danger if these models were applied in practice without knowledge of the sublanguage of the law. In this paper, we raise the issue and propose a trivial solution by introducing BERTLaw a legal sublanguage pretrained model. The paper's experiments demonstrate the superior effectiveness of the method compared to the baseline pretrained model