{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.20204", "kind": "paper", "url": "https://arxiv.org/abs/2608.20204"}], "authors": ["Yejin Bang", "Kirsty Fielding", "Brandan Oliver", "Brian Birke", "Nabeel Seedat", "Andrew M. Bean"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.20204", "source_url": "https://arxiv.org/abs/2608.20204"}, "categories": ["General AI", "Language & Knowledge", "cs.AI"], "description": {"en": "Evaluates LLMs on legal contract scrubbing tasks using hand-crafted contracts covering error categories like defined term misuse and inconsistent language."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.20204", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.20204", "title": "ContractScrub"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Thomson Reuters (United Kingdom)", "Thomson Reuters Foundation", "Imperial College London"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2608.20204", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2608.20204", "source_url": "https://openalex.org/W7203949388"}, "key": "claire-radar:2608.20204", "modality": null, "name": "ContractScrub", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_contractscrub_bad94249", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "15c4ee0ad2523559779de17556a9f2f94c55fb5056eca6f33bb0a22ddd4f50e6", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-08-25T23:23:46.758372Z", "source_url": "https://arxiv.org/abs/2608.20204"}, "publisher": null, "released": "2026-08-20", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.20204", "source_url": "https://arxiv.org/abs/2608.20204"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-20204", "source": "claire_radar", "source_benchmark_id": "2608.20204", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 0, "hfPaperUrl": "https://huggingface.co/papers/2608.20204", "observedAt": "2026-10-04T03:02:42.754940Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-09-30", "state": "stale"}}, "status": "stale"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "The domain-specific legal focus and the clear performance gap reported may attract attention from both legal AI and LLM evaluation communities.", "score": 58}, "capabilities": [], "confidence": "High", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-25T23:23:46.758372Z", "inputHash": "629e15406551a636fd2951b8ab72bbca614349821e0b32d5bce32d9feb4aca4f", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "ContractScrub: A benchmark for final review of legal contracts", "canonicalNameSource": "paper_title", "decisionReason": "The paper introduces a specific evaluation object with defined error categories and a clear scoring metric (macro average recall), and the arxiv report serves as a public reference.", "model": "deepseek-v4-pro", "reviewedAt": "2026-08-25T23:23:46.758372Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates LLMs on legal contract scrubbing tasks using hand-crafted contracts covering error categories like defined term misuse and inconsistent language.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["exact coined title identity tied to benchmark evidence", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We introduce ContractScrub, the first benchmark designed to evaluate contract scrubbing capabilities, comprising contracts hand-crafted by experienced lawyers over diverse error categories such as misuse of defined terms, incorrect references, and inconsistent language."}, "familyId": "bmf_0d037e92ca7f", "firstSeenAt": "2026-08-23", "heat": null, "id": "bm_contractscrub_bad94249", "indexedAt": "2026-08-23T06:55:50.392548Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2608.20204", "pdf": "https://arxiv.org/pdf/2608.20204", "project": null, "report": "https://arxiv.org/abs/2608.20204"}, "metrics": [], "motivation": "Legal work, with its heavy reliance on processing large amounts of text, is often considered one of the domains most exposed to the use of LLMs.", "name": "ContractScrub", "oneLine": "Evaluates LLMs on legal contract scrubbing tasks using hand-crafted contracts covering error categories like defined term misuse and inconsistent language.", "paperTitle": "ContractScrub: A benchmark for final review of legal contracts", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.20204", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-23T06:55:50.392548Z"}, "publications": [], "ranking": {"90d": {"confidence": "Low", "coverage": 0.15, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": null, "value": null}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.0, "value": 0}}, "confidence": "Low", "coverage": 0.15, "rank": null, "score": 43}, "method": "attention-ranking-v13", "rank": null, "score": 43}}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-08-20", "source": {"authors": ["Yejin Bang", "Kirsty Fielding", "Brandan Oliver", "Brian Birke", "Nabeel Seedat", "Andrew M. Bean"], "categories": ["cs.AI", "cs.CL"], "id": "2608.20204", "title": "ContractScrub: A benchmark for final review of legal contracts", "type": "arxiv", "url": "https://arxiv.org/abs/2608.20204"}, "sourceUpdatedAt": "2026-08-20T16:01:42Z", "topics": ["cs.AI"], "venueAttempts": [], "whyItMatters": "Provides the first formal evaluation of contract scrubbing, revealing that frontier models perform surprisingly poorly on this specific legal review task despite strong general benchmarks."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
