{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.08932", "kind": "paper", "url": "https://arxiv.org/abs/2606.08932"}], "authors": ["Jian Chen", "Siyuan Li", "Chucheng Wan", "Zixuan Yuan"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.08932", "source_url": "https://arxiv.org/abs/2606.08932"}, "categories": ["Language & Knowledge", "Software & AI Compute", "cs.CL"], "description": {"en": "NormBench evaluates defeasible scope parsing in legal texts using Span-Grounded Deontic Trees, with 2,290 provisions across multiple languages. It focuses on identifying clause overrides and includes whole-tree fidelity metrics."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.08932", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.08932", "title": "NormBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["The Hong Kong University of Science and Technology (Guangzhou)", "Sun Yat-sen University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.08932", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.08932", "source_url": "https://openalex.org/W7164053633"}, "key": "claire-radar:2606.08932", "modality": null, "name": "NormBench", "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": "unknown", "origin_record_id": "bm_normbench_a8bf7215", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "4deefe4baf8d266a3dd4943ccefbb4bbd3b6904189a29140c47c2d14700b2758", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2606.08932"}, "publisher": null, "released": "2026-06-08", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.08932", "source_url": "https://arxiv.org/abs/2606.08932"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-08932", "source": "claire_radar", "source_benchmark_id": "2606.08932", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Mixed", "applicationDomains": ["Software & AI Compute"], "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": null, "hfPaperUrl": "https://huggingface.co/papers/2606.08932", "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", "state": "unavailable"}}, "status": "unavailable"}, "capabilities": [], "confidence": "High", "construction": "Aggregate Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "56dbc27679f4534c96d8a9c38161aeb241021744fead820c3614890513af863d", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "NormBench evaluates defeasible scope parsing in legal texts using Span-Grounded Deontic Trees, with 2,290 provisions across multiple languages. It focuses on identifying clause overrides and includes whole-tree fidelity metrics.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact named benchmark artifact released in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "To diagnose and mitigate SSO, we introduce NormBench, a benchmark of 2,290 provisions spanning Chinese (laws and local policies), English (U.S."}, "familyId": "bmf_8a7f98f06ad7", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_normbench_a8bf7215", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Software & Cloud"], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2606.08932", "paper": "https://arxiv.org/abs/2606.08932", "pdf": "https://arxiv.org/pdf/2606.08932", "project": null, "report": "https://arxiv.org/abs/2606.08932"}, "metrics": [], "motivation": "Rule-following agents tasked with executing policies and regulations often fail via Silent Scope Omission (SSO): a model applies a general rule but silently drops nested exceptions or counter-exceptions, producing outputs that appear compliant yet break on important edge cases.", "name": "NormBench", "oneLine": "NormBench evaluates defeasible scope parsing in legal texts using Span-Grounded Deontic Trees, with 2,290 provisions across multiple languages. It focuses on identifying clause overrides and includes whole-tree fidelity metrics.", "paperTitle": "From Statute to Control Flow: Span-Grounded Deontic Trees for Defeasible Scope Parsing", "primaryDomain": "Software & AI Compute", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.08932", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-06-08", "source": {"authors": ["Jian Chen", "Siyuan Li", "Chucheng Wan", "Zixuan Yuan"], "categories": ["cs.CL", "cs.AI", "cs.CE"], "id": "2606.08932", "title": "From Statute to Control Flow: Span-Grounded Deontic Trees for Defeasible Scope Parsing", "type": "arxiv", "url": "https://arxiv.org/abs/2606.08932"}, "sourceUpdatedAt": "2026-06-08T02:17:53Z", "topics": ["cs.CL"], "venueAttempts": [], "whyItMatters": "Silent Scope Omission is a critical failure in rule-following agents. NormBench provides a diagnostic benchmark to identify structural omissions and improve statutory understanding in LLMs."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
