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

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

18 Apr 2021arXiv:2104.08671archive 2025-07-28

Lucia Zheng, Neel Guha, Brandon R. Anderson, Peter Henderson, Daniel E. Ho

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.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 9 code samples harvested from 2 repositories linked to this paper: 6 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: official repository: 1 sample from 1 repository, 1 ran (1 with no instrument failure; 0 where Syntology's instrument failed); community (archive-listed): 8 samples from 1 repository, 5 ran (5 with no instrument failure; 0 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-24 and 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

reglab/casehold officialmentioned in papermentioned on GitHubpytorchApache-2.0Syntology: 1 sample, 1 ran (0 instrument-failed) · 0 unverified report
trusthlt/privacy-legal-nlp-lm mentioned on GitHubjaxApache-2.0Syntology: 8 samples, 5 ran (0 instrument-failed) · 3 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

9 samples harvested: 6 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

6ran
3unverified

Licence: 0 of the 9 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 2 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

reglab/casehold official repository 1 sample, 1 ran (0 instrument-failed) · 0 unverified
convert_examples_to_features in 134 papers multiple_choice/utils_multiple_choice.py ran Apache-2.0 (permissive) · 68377121bfc285a9 · report fixture: signature examples:list/1/InputExample, label_list:list/1/str, max_length:int/0/int, tokenizer:callable/0/? · returned list · level recorded 2026-09-25
trusthlt/privacy-legal-nlp-lm community (archive-listed) 8 samples, 5 ran (0 instrument-failed) · 3 unverified
apply_threshs downstream_tasks/utils.py ran fingerprinted Apache-2.0 (permissive) · e1f72af0ac263f37 · report fixture: signature probas:ndarray/2/float, threshs:ndarray/1/float · returned ndarray · level recorded 2026-09-24
create_instance_from_document data_prepare/utils.py ran Apache-2.0 (permissive) · 36a781e6d5afa906 · report fixture: signature document:list/1/list, doc_idx:int/0/int, datasets:list/1/dict, version:str/0/str, max_seq_length:int/0/int, short_seq_prob:float/0/float, special_tokens:dict/0/? · returned dict · level recorded 2026-09-24
equal pretrain/utils.py ran Apache-2.0 (permissive) · 45127374f48c9723 · report fixture: signature x1:float/0/float, x2:float/0/float, diff:float/0/float · returned bool · level recorded 2026-09-24
generate_random_next data_prepare/utils.py ran Apache-2.0 (permissive) · efbd156ad839d2bf · report fixture: signature datasets:list/1/dict, doc_idx:int/0/int, target_b_length:int/0/int · returned list · level recorded 2026-09-24
multihot downstream_tasks/data_loader.py ran Apache-2.0 (permissive) · 851b9497b3ad4904 · report fixture: signature labels:list/1/str, label_map:dict/0/? · returned ndarray · level recorded 2026-09-24
build_dataset in 151 papers downstream_tasks/data_loader.py unverified Apache-2.0 (permissive) · f252d20c754d3bcf · report
split_corpus downstream_tasks/data_loader.py unverified Apache-2.0 (permissive) · 14ea19820452ef31 · report
tune_threshs downstream_tasks/utils.py unverified Apache-2.0 (permissive) · a58a2a6534b81101 · report

Tasks

Multiple-choiceQuestion AnsweringSelf-Supervised LearningSpecificityText Classification

Datasets

Introduced by this paper, per the archive.

CaseHOLDOverruling

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Question Answering CaseHOLD Custom Legal-BERT Macro F1 (10-fold) 69.5 #1 of 3 Archive leaderboard report
Question Answering CaseHOLD Legal-BERT Macro F1 (10-fold) 68.0 #2 of 3 Archive leaderboard report
Question Answering CaseHOLD BERT Macro F1 (10-fold) 61.3 #3 of 3 Archive leaderboard report
Text Classification Overruling Custom Legal-BERT F1(10-fold) 97.4 #1 of 3 Archive leaderboard report
Text Classification Overruling Legal-BERT F1(10-fold) 96.3 #2 of 3 Archive leaderboard report
Text Classification Overruling BERT F1(10-fold) 95.8 #3 of 3 Archive leaderboard report
Text Classification Terms of Service Custom Legal-BERT F1(10-fold) 78.7 #1 of 3 Archive leaderboard report
Text Classification Terms of Service Legal-BERT F1(10-fold) 75.0 #2 of 3 Archive leaderboard report
Text Classification Terms of Service BERT F1(10-fold) 72.2 #3 of 3 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Absolute Position EncodingsAdamAttentionBPEBiLSTMDense ConnectionsDropoutLSTMLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSigmoid ActivationSoftmaxTanh ActivationTransformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections