{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.20634", "kind": "paper", "url": "https://arxiv.org/abs/2606.20634"}, {"id": "gh:agent-runtime-evidence/decision-evidence-benchmark", "kind": "repo", "url": "https://github.com/agent-runtime-evidence/decision-evidence-benchmark"}, {"id": "hf:dev404ai/demm-bench", "kind": "dataset", "url": "https://huggingface.co/datasets/dev404ai/DEMM-Bench"}], "authors": ["Oleg Solozobov"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.20634", "source_url": "https://arxiv.org/abs/2606.20634"}, "categories": ["Agents", "General AI", "Language & Knowledge"], "description": {"en": "A cross-regime benchmark for testing whether agent-runtime records are sufficient to reconstruct decision-level governance properties under deterministic degradation conditions."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.20634", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.20634", "title": "DEMM-Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2606.20634", "modality": null, "name": "DEMM-Bench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-09-25T04:40:22Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_demm-bench_db258a39", "origin_source": "arxiv", "recognition_confidence": "0.6", "record_sha256": "bcddd5dc286b9774d7c92a1be1ea86e8d5f2018cb7ac1e7ce40ddd6dc4f49a35", "relation": "introduces", "review_model": "codex-cli-default", "review_state": "ai-reviewed", "reviewed_at": "2026-09-24T18:11:26.867488Z", "source_url": "https://arxiv.org/abs/2606.20634"}, "publisher": null, "released": "2026-05-30", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.20634", "source_url": "https://arxiv.org/abs/2606.20634"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-20634", "source": "claire_radar", "source_benchmark_id": "2606.20634", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attentionForecast": {"confidence": "Low", "horizon": "7d", "reason": "Agent governance is an emerging area with potential specialist interest, while the benchmark's narrow scope may constrain near-term reach.", "score": 48}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-24T18:11:26.867488Z", "inputHash": "4b306f403cb6253bbbd3e5f993c6d8554a026ea8cbfcf38061259c96f6a2b758", "model": "codex-cli-default"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "DEMM-Bench: A Cross-Regime Benchmark for Agent-Runtime Governance-Evidence Sufficiency", "canonicalNameSource": "paper_title", "decisionReason": "The source supplies a formal benchmark name, a defined task, a comparable accuracy measure, deterministic test conditions, and licensed public dataset and code paths.", "evaluationEvidence": {"reuse": "Dataset (CC-BY-4.0) and code (Apache-2.0): Zenodo doi:10.5281/zenodo.20426092 and Hugging Face https://huggingface.co/datasets/dev404ai/DEMM-Bench; code at https://github.com/agent-runtime-evidence/decision-evidence-benchmark", "scoring": "56.25% mean Property Sufficiency Accuracy", "task": "measures whether records across eight evidence regimes are sufficient to reconstruct decision-level properties rather than merely present"}, "model": "codex-cli-default", "reviewedAt": "2026-09-24T18:11:26.867488Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "A cross-regime benchmark for testing whether agent-runtime records are sufficient to reconstruct decision-level governance properties under deterministic degradation conditions.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "public artifact URL"], "snippet": "DEMM-Bench is a cross-regime benchmark for agent-runtime governance-evidence sufficiency, grounded in the Decision Evidence Maturity Model (DEMM): it measures whether records across eight evidence regimes are sufficient to reconstruct decision-level properties rather than merely present."}, "familyId": "bmf_fd2f536eb900", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_demm-bench_db258a39", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": "https://github.com/agent-runtime-evidence/decision-evidence-benchmark", "data": "https://huggingface.co/datasets/dev404ai/DEMM-Bench", "paper": "https://arxiv.org/abs/2606.20634", "pdf": "https://arxiv.org/pdf/2606.20634", "project": null, "report": "https://arxiv.org/abs/2606.20634"}, "metrics": [], "motivation": "Agent-runtime systems emit traces, ledgers, provenance graphs, policy logs, delegation tokens, cache events, and tool-firewall records, but those containers do not necessarily answer governance questions about a specific decision.", "name": "DEMM-Bench", "oneLine": "A cross-regime benchmark for testing whether agent-runtime records are sufficient to reconstruct decision-level governance properties under deterministic degradation conditions.", "paperTitle": "DEMM-Bench: A Cross-Regime Benchmark for Agent-Runtime Governance-Evidence Sufficiency", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.20634", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "publishers": [{"name": "agent-runtime-evidence", "organizationType": "community", "role": "benchmark-publisher", "sourceUrl": "https://github.com/agent-runtime-evidence/decision-evidence-benchmark"}], "readiness": "Runnable", "recognitionConfidence": 0.6, "relation": "introduces", "releasedAt": "2026-05-30", "reviewLifecycle": {"state": "pending"}, "source": {"authors": ["Oleg Solozobov"], "categories": ["cs.AI", "cs.CY"], "id": "2606.20634", "title": "DEMM-Bench: A Cross-Regime Benchmark for Agent-Runtime Governance-Evidence Sufficiency", "type": "arxiv", "url": "https://arxiv.org/abs/2606.20634"}, "sourceUpdatedAt": "2026-05-30T09:36:29Z", "topics": ["Agents"], "venueAttempts": [], "whyItMatters": "It provides a reproducible way to distinguish the presence of runtime records from their actual evidentiary sufficiency."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
