{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2607.02825", "kind": "paper", "url": "https://arxiv.org/abs/2607.02825"}], "authors": ["Norbert Sandor Szolnoki", "Gabor Antal"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2607.02825", "source_url": "https://arxiv.org/abs/2607.02825"}, "categories": ["Cybersecurity", "Vision & 3D", "cs.CR"], "description": {"en": "JavaVulBench is a benchmark dataset and evaluation harness for Java vulnerability detection. The dataset includes about 30,600 Java methods spanning 1,740 CVEs and 700+ projects, with method and line labels, publication dates, and five split strategies. The harness provides a unified schema across multiple backends and includes a contamination audit."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2607.02825", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2607.02825", "title": "JavaVulBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["University of Szeged"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2607.02825", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2607.02825", "source_url": "https://openalex.org/W7167620052"}, "key": "claire-radar:2607.02825", "modality": null, "name": "JavaVulBench", "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_javavulbench_93d353dc", "origin_source": "arxiv", "recognition_confidence": "0.95", "record_sha256": "077c83b501d581be197486e3bd25e02f6ba700de2187cdb9abd06c2b4f476149", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2607.02825"}, "publisher": null, "released": "2026-07-02", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2607.02825", "source_url": "https://arxiv.org/abs/2607.02825"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2607-02825", "source": "claire_radar", "source_benchmark_id": "2607.02825", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["Cybersecurity"], "area": "Vision & 3D", "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/2607.02825", "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": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "cb72eb271cd2c494f1b7ccfd68ed2000fda3b7e3a003465535675ada1e918fb7", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "JavaVulBench is a benchmark dataset and evaluation harness for Java vulnerability detection. The dataset includes about 30,600 Java methods spanning 1,740 CVEs and 700+ projects, with method and line labels, publication dates, and five split strategies. The harness provides a unified schema across multiple backends and includes a contamination audit.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "public artifact URL"], "snippet": "We release \\textsc{JavaVulBench}, a benchmark dataset and evaluation harness for Java vulnerability detection."}, "familyId": "bmf_c5f6848b3970", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_javavulbench_93d353dc", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Cybersecurity"], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2607.02825", "paper": "https://arxiv.org/abs/2607.02825", "pdf": "https://arxiv.org/pdf/2607.02825", "project": "https://www.youtube.com/watch?v=nMTX\\_hqkuoM", "report": "https://arxiv.org/abs/2607.02825"}, "metrics": [], "motivation": "We release \\textsc{JavaVulBench}, a benchmark dataset and evaluation harness for Java vulnerability detection.", "name": "JavaVulBench", "oneLine": "JavaVulBench is a benchmark dataset and evaluation harness for Java vulnerability detection. The dataset includes about 30,600 Java methods spanning 1,740 CVEs and 700+ projects, with method and line labels, publication dates, and five split strategies. The harness provides a unified schema across multiple backends and includes a contamination audit.", "paperTitle": "JavaVulBench: A Java Vulnerability Benchmark with Realistic Splits, a Unified Multi-Backend Harness, and a Leakage-Aware Evaluation Mode", "primaryDomain": "Cybersecurity", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2607.02825", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {}, "readiness": "Inspectable", "recognitionConfidence": 0.95, "relation": "introduces", "releasedAt": "2026-07-02", "source": {"authors": ["Norbert Sandor Szolnoki", "Gabor Antal"], "categories": ["cs.CR", "cs.AI"], "id": "2607.02825", "title": "JavaVulBench: A Java Vulnerability Benchmark with Realistic Splits, a Unified Multi-Backend Harness, and a Leakage-Aware Evaluation Mode", "type": "arxiv", "url": "https://arxiv.org/abs/2607.02825"}, "sourceUpdatedAt": "2026-07-02T23:32:43Z", "topics": ["cs.CR"], "venueAttempts": [], "whyItMatters": "It addresses the need for realistic and leakage-aware evaluation of vulnerability detection models, offering multiple split strategies and contamination audits to separate genuinely unseen CVEs from memorized ones."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
