{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/lexglue-a-benchmark-dataset-for-legal","title":"LexGLUE: A Benchmark Dataset for Legal Language Understanding in English","arxiv_id":"2110.00976","date":"2021-10-03","proceeding":"ACL 2022 5","authors":["Ilias Chalkidis","Abhik Jana","Dirk Hartung","Michael Bommarito","Ion Androutsopoulos","Daniel Martin Katz","Nikolaos Aletras"],"abstract":"Laws and their interpretations, legal arguments and agreements\\ are typically expressed in writing, leading to the production of vast corpora of legal text. Their analysis, which is at the center of legal practice, becomes increasingly elaborate as these collections grow in size. Natural language understanding (NLU) technologies can be a valuable tool to support legal practitioners in these endeavors. Their usefulness, however, largely depends on whether current state-of-the-art models can generalize across various tasks in the legal domain. To answer this currently open question, we introduce the Legal General Language Understanding Evaluation (LexGLUE) benchmark, a collection of datasets for evaluating model performance across a diverse set of legal NLU tasks in a standardized way. We also provide an evaluation and analysis of several generic and legal-oriented models demonstrating that the latter consistently offer performance improvements across multiple tasks.","url_abs":"https://arxiv.org/abs/2110.00976v4","url_pdf":"https://arxiv.org/pdf/2110.00976v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"lexglue-a-benchmark-dataset-for-legal","repo_url":"https://github.com/coastalcph/lex-glue","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"multi-label-classification","task_name":"Multi-Label Classification"},{"task_slug":"multi-class-classification","task_name":"Multi-class Classification"},{"task_slug":"multiple-choice-qa","task_name":"Multiple Choice Question Answering (MCQA)"},{"task_slug":"natural-language-understanding","task_name":"Natural Language Understanding"},{"task_slug":"open-question","task_name":"Open-Ended Question Answering"}],"methods":[],"datasets_introduced":[{"slug":"lexglue","name":"LexGLUE","full_name":""}],"methods_introduced":[],"results":[{"leaderboard":"/sota/natural-language-understanding-on-lexglue","task":"Natural Language Understanding","dataset":"LexGLUE","model":"BERT","rank_in_archive_order":1,"of":8,"metrics":{"CaseHOLD":"70.7","ECtHR Task A":"71.4 / 64.0","ECtHR Task B":"87.6 / 77.8","EUR-LEX":"71.6 / 55.6","LEDGAR":"87.7 / 82.2","SCOTUS":"70.5 / 60.9","UNFAIR-ToS":"87.5 / 81.0"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-understanding-on-lexglue","task":"Natural Language Understanding","dataset":"LexGLUE","model":"Legal-BERT","rank_in_archive_order":2,"of":8,"metrics":{"CaseHOLD":"75.1","ECtHR Task A":"71.2 / 64.6","ECtHR Task B":"88.0 / 77.2","EUR-LEX":"72.2 / 56.2","LEDGAR":"88.1 / 82.7","SCOTUS":"76.2 / 65.8","UNFAIR-ToS":"88.6 / 82.3"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-understanding-on-lexglue","task":"Natural Language Understanding","dataset":"LexGLUE","model":"CaseLaw-BERT","rank_in_archive_order":3,"of":8,"metrics":{"CaseHOLD":"75.6","ECtHR Task A":"71.2 / 64.2","ECtHR Task B":"88.0 / 77.5","EUR-LEX":"71.0 / 55.9","LEDGAR":"88.0 / 82.3","SCOTUS":"76.4 / 66.2","UNFAIR-ToS":"88.3 / 81.0"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-understanding-on-lexglue","task":"Natural Language Understanding","dataset":"LexGLUE","model":"BigBird","rank_in_archive_order":4,"of":8,"metrics":{"CaseHOLD":"70.4","ECtHR Task A":"70.5 / 63.8","ECtHR Task B":"88.1 / 76.6","EUR-LEX":"71.8 / 56.6","LEDGAR":"87.7 / 82.1","SCOTUS":"71.7 / 61.4","UNFAIR-ToS":"87.7 / 80.2"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-understanding-on-lexglue","task":"Natural Language Understanding","dataset":"LexGLUE","model":"Longformer","rank_in_archive_order":5,"of":8,"metrics":{"CaseHOLD":"72.0","ECtHR Task A":"69.6 / 62.4","ECtHR Task B":"88.0 / 77.8","EUR-LEX":"71.9 / 56.7","LEDGAR":"87.7 / 82.3","SCOTUS":"72.2 / 62.5","UNFAIR-ToS":"87.7 / 80.1"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-understanding-on-lexglue","task":"Natural Language Understanding","dataset":"LexGLUE","model":"RoBERTa","rank_in_archive_order":6,"of":8,"metrics":{"CaseHOLD":"71.7","ECtHR Task A":"69.5 / 60.7","ECtHR Task B":"87.2 / 77.3","EUR-LEX":"71.8 / 57.5","LEDGAR":"87.9 / 82.1","SCOTUS":"70.8 / 61.2","UNFAIR-ToS":"87.7 / 81.5"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-understanding-on-lexglue","task":"Natural Language Understanding","dataset":"LexGLUE","model":"DeBERTa","rank_in_archive_order":7,"of":8,"metrics":{"CaseHOLD":"72.1","ECtHR Task A":"69.1 / 61.2","ECtHR Task B":"87.4 / 77.3","EUR-LEX":"72.3 / 57.2","LEDGAR":"87.9 / 82.0","SCOTUS":"70.0 / 60.0","UNFAIR-ToS":"87.2 / 78.8"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2110.00976","atlas_url":"https://app.syntology.ai/?focus=2110.00976","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2110.00976"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/coastalcph/lex-glue","reach":{"status":"ok"}}],"summary":{"unverified":1},"by_repo_kind":{},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"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"},"n_pointer_only_for_licence":1,"samples":[{"code_sha256_prefix":"05732f1a5126672d","entry":"get_mask","repo":null,"repo_kind":null,"path":null,"file_url":null,"link_basis":"identical_code_first_harvested_elsewhere","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":null,"inline_ok":false,"mcp_get_code":{"code_sha256":"05732f1a5126672d"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}