# When Does Pretraining Help? Assessing Self-Supervised Learning for Law and the CaseHOLD Dataset

arXiv: https://arxiv.org/abs/2104.08671 · HTML: https://mlsanity.com/paper/2104.08671/

## Key facts

- Cited by 201 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 55 references that are in the map.
- First posted on arXiv 18 Apr 2021.

## Builds on (most cited first, 27 of 27)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding](https://mlsanity.com/paper/1804.07461/) (2018)
- [Efficient Estimation of Word Representations in Vector Space](https://mlsanity.com/paper/1301.3781/) (2013)
- [SQuAD: 100,000+ Questions for Machine Comprehension of Text](https://mlsanity.com/paper/1606.05250/) (2016)
- [SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing](https://mlsanity.com/paper/1808.06226/) (2018)
- [BioBERT: a pre-trained biomedical language representation model for biomedical text mining](https://mlsanity.com/paper/1901.08746/) (2019)
- [Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation](https://mlsanity.com/paper/1609.08144/) (2016)
- [SciBERT: A Pretrained Language Model for Scientific Text](https://mlsanity.com/paper/1903.10676/) (2019)
- [SpanBERT: Improving Pre-training by Representing and Predicting Spans](https://mlsanity.com/paper/1907.10529/) (2019)
- [A Primer in BERTology: What we know about how BERT works](https://mlsanity.com/paper/2002.12327/) (2020)
- [Aligning AI With Shared Human Values](https://mlsanity.com/paper/2008.02275/) (2020)
- [CoQA: A Conversational Question Answering Challenge](https://mlsanity.com/paper/1808.07042/) (2018)
- [How Good is Your Tokenizer? On the Monolingual Performance of Multilingual Language Models](https://mlsanity.com/paper/2012.15613/) (2020)
- [ERNIE: Enhanced Representation through Knowledge Integration](https://mlsanity.com/paper/1904.09223/) (2019)
- [Neural Legal Judgment Prediction in English](https://mlsanity.com/paper/1906.02059/) (2019)
- [emrQA: A Large Corpus for Question Answering on Electronic Medical Records](https://mlsanity.com/paper/1809.00732/) (2018)
- [How Does NLP Benefit Legal System: A Summary of Legal Artificial Intelligence](https://mlsanity.com/paper/2004.12158/) (2020)
- [JEC-QA: A Legal-Domain Question Answering Dataset](https://mlsanity.com/paper/1911.12011/) (2019)
- [CLAUDETTE: an Automated Detector of Potentially Unfair Clauses in Online Terms of Service](https://mlsanity.com/paper/1805.01217/) (2018)
- [UnNatural Language Inference](https://mlsanity.com/paper/2101.00010/) (2020)
- [Exploring the Use of Text Classification in the Legal Domain](https://mlsanity.com/paper/1710.09306/) (2017)
- [LexNLP: Natural language processing and information extraction for legal and regulatory texts](https://mlsanity.com/paper/1806.03688/) (2018)
- [BERT Goes to Law School: Quantifying the Competitive Advantage of Access to Large Legal Corpora in Contract Understanding](https://mlsanity.com/paper/1911.00473/) (2019)
- [Neural Mask Generator: Learning to Generate Adaptive Word Maskings for Language Model Adaptation](https://mlsanity.com/paper/2010.02705/) (2020)
- [Neural Contract Element Extraction Revisited: Letters from Sesame Street](https://mlsanity.com/paper/2101.04355/) (2021)

## Built on by (most cited first, 30 of 201)

- [FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance](https://mlsanity.com/paper/2305.05176/) (2023)
- [LegalBench: A Collaboratively Built Benchmark for Measuring Legal Reasoning in Large Language Models](https://mlsanity.com/paper/2308.11462/) (2023)
- [LexGLUE: A Benchmark Dataset for Legal Language Understanding in English](https://mlsanity.com/paper/2110.00976/) (2021)
- [Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Tools](https://mlsanity.com/paper/2405.20362/) (2024)
- [ContractNLI: A Dataset for Document-level Natural Language Inference for Contracts](https://mlsanity.com/paper/2110.01799/) (2021)
- [Pile of Law: Learning Responsible Data Filtering from the Law and a 256GB Open-Source Legal Dataset](https://mlsanity.com/paper/2207.00220/) (2022)
- [The CoT Collection: Improving Zero-shot and Few-shot Learning of Language Models via Chain-of-Thought Fine-Tuning](https://mlsanity.com/paper/2305.14045/) (2023)
- [HuatuoGPT-II, One-stage Training for Medical Adaption of LLMs](https://mlsanity.com/paper/2311.09774/) (2023)
- [LLM2LLM: Boosting LLMs with Novel Iterative Data Enhancement](https://mlsanity.com/paper/2403.15042/) (2024)
- [Can multiple-choice questions really be useful in detecting the abilities of LLMs?](https://mlsanity.com/paper/2403.17752/) (2024)
- [LEXTREME: A Multi-Lingual and Multi-Task Benchmark for the Legal Domain](https://mlsanity.com/paper/2301.13126/) (2023)
- [(A)I Am Not a Lawyer, But...: Engaging Legal Experts towards Responsible LLM Policies for Legal Advice](https://mlsanity.com/paper/2402.01864/) (2024)
- [MultiLegalPile: A 689GB Multilingual Legal Corpus](https://mlsanity.com/paper/2306.02069/) (2023)
- [Transferability in Deep Learning: A Survey](https://mlsanity.com/paper/2201.05867/) (2022)
- [LeXFiles and LegalLAMA: Facilitating English Multinational Legal Language Model Development](https://mlsanity.com/paper/2305.07507/) (2023)
- [Lawyer LLaMA Technical Report](https://mlsanity.com/paper/2305.15062/) (2023)
- [A Short Survey of Viewing Large Language Models in Legal Aspect](https://mlsanity.com/paper/2303.09136/) (2023)
- [A Survey on Legal Judgment Prediction: Datasets, Metrics, Models and Challenges](https://mlsanity.com/paper/2204.04859/) (2022)
- [Adapting Large Language Models to Domains via Reading Comprehension](https://mlsanity.com/paper/2309.09530/) (2023)
- [Multi-LexSum: Real-World Summaries of Civil Rights Lawsuits at Multiple Granularities](https://mlsanity.com/paper/2206.10883/) (2022)
- [Training Data Extraction From Pre-trained Language Models: A Survey](https://mlsanity.com/paper/2305.16157/) (2023)
- [RAFT: A Real-World Few-Shot Text Classification Benchmark](https://mlsanity.com/paper/2109.14076/) (2021)
- [Natural Language Processing for the Legal Domain: A Survey of Tasks, Datasets, Models, and Challenges](https://mlsanity.com/paper/2410.21306/) (2024)
- [Empower Large Language Model to Perform Better on Industrial Domain-Specific Question Answering](https://mlsanity.com/paper/2305.11541/) (2023)
- [Natural Language Processing in the Legal Domain](https://mlsanity.com/paper/2302.12039/) (2023)
- [A Reasoning-Focused Legal Retrieval Benchmark](https://mlsanity.com/paper/2505.03970/) (2025)
- [Mining Legal Arguments in Court Decisions](https://mlsanity.com/paper/2208.06178/) (2022)
- [MAUD: An Expert-Annotated Legal NLP Dataset for Merger Agreement Understanding](https://mlsanity.com/paper/2301.00876/) (2023)
- [ChatGPT may Pass the Bar Exam soon, but has a Long Way to Go for the LexGLUE benchmark](https://mlsanity.com/paper/2304.12202/) (2023)
- [Multi-Task Inference: Can Large Language Models Follow Multiple Instructions at Once?](https://mlsanity.com/paper/2402.11597/) (2024)

## Abstract (arXiv)

While self-supervised learning has made rapid advances in natural language processing, it remains unclear when researchers should engage in resource-intensive domain-specific pretraining (domain pretraining). The law, puzzlingly, has yielded few documented instances of substantial gains to domain pretraining in spite of the fact that legal language is widely seen to be unique. We hypothesize that these existing results stem from the fact that existing legal NLP tasks are too easy and fail to meet conditions for when domain pretraining can help. To address this, we first present CaseHOLD (Case Holdings On Legal Decisions), a new dataset comprised of over 53,000+ multiple choice questions to identify the relevant holding of a cited case. This dataset presents a fundamental task to lawyers and is both legally meaningful and difficult from an NLP perspective (F1 of 0.4 with a BiLSTM baseline). Second, we assess performance gains on CaseHOLD and existing legal NLP datasets. While a Transformer architecture (BERT) pretrained on a general corpus (Google Books and Wikipedia) improves performance, domain pretraining (using corpus of approximately 3.5M decisions across all courts in the U.S. that is larger than BERT's) with a custom legal vocabulary exhibits the most substantial performance gains with CaseHOLD (gain of 7.2% on F1, representing a 12% improvement on BERT) and consistent performance gains across two other legal tasks. Third, we show that domain pretraining may be warranted when the task exhibits sufficient similarity to the pretraining corpus: the level of performance increase in three legal tasks was directly tied to the domain specificity of the task. Our findings inform when researchers should engage resource-intensive pretraining and show that Transformer-based architectures, too, learn embeddings suggestive of distinct legal language.