{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.05212", "kind": "paper", "url": "https://arxiv.org/abs/2608.05212"}], "categories": ["General AI", "Language & Knowledge", "cs.AI"], "description": {"en": "SearchAuditBench evaluates LLM auditors on localizing, attributing, and repairing failures in long-horizon search agent trajectories using expert-annotated error steps, root causes, and repair rubrics."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.05212", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.05212", "title": "SearchAuditBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2608.05212", "modality": null, "name": "SearchAuditBench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "Medium", "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_searchauditbench_2dd8deb6", "origin_source": "arxiv", "recognition_confidence": "0.55", "record_sha256": "301931b3af77167a5d5299e091296437825d74acd371583870f176992cc64cb1", "relation": "evaluates_only", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-09-08T08:57:27.463024Z", "source_url": "https://arxiv.org/abs/2608.05212"}, "publisher": null, "released": "2026-08-05", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.05212", "source_url": "https://arxiv.org/abs/2608.05212"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-05212", "source": "claire_radar", "source_benchmark_id": "2608.05212", "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": null, "hfPaperUrl": "https://huggingface.co/papers/2608.05212", "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"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "Auditing long-horizon search agents is a niche but timely problem; the benchmark's scale and expert annotations may draw interest from agent evaluation researchers.", "score": 55}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-08T08:57:27.463024Z", "inputHash": "c9297f61205ca60b95d6a51cdaf7922ead7c5fdb3215b72df563548cd5103e51", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "We therefore introduce SearchAuditBench, a benchmark that evaluates whether LLM auditors can localize, attribute, and repair these failures", "canonicalNameSource": "abstract", "decisionReason": "The benchmark is formally named in the abstract, includes a defined protocol with expert annotations and grading rubrics, and is intended for objective comparison of auditing models.", "model": "deepseek-v4-pro", "reviewedAt": "2026-09-08T08:57:27.463024Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "SearchAuditBench evaluates LLM auditors on localizing, attributing, and repairing failures in long-horizon search agent trajectories using expert-annotated error steps, root causes, and repair rubrics.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["exact named benchmark artifact released in abstract", "evaluation protocol evidence", "named entity is a model, method, framework, or dataset used in evaluation"], "snippet": "We therefore introduce SearchAuditBench, a benchmark that evaluates whether LLM auditors can localize, attribute, and repair these failures, thereby reducing the human burden."}, "familyId": "bmf_daecf0ea3bc7", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_searchauditbench_2dd8deb6", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2608.05212", "pdf": "https://arxiv.org/pdf/2608.05212", "project": null, "report": "https://arxiv.org/abs/2608.05212"}, "metrics": [], "motivation": "Deep search agents tackle challenging questions through long-horizon web interactions, a process that is both complex and fragile: small reasoning errors may propagate through long, noisy trajectories into fluent but incorrect answers.", "name": "SearchAuditBench", "oneLine": "SearchAuditBench evaluates LLM auditors on localizing, attributing, and repairing failures in long-horizon search agent trajectories using expert-annotated error steps, root causes, and repair rubrics.", "paperTitle": "SearchAuditor: Auditing and Attributing Failures in Long-Horizon Search Agents", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.05212", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "ranking": {"90d": {"confidence": "Low", "coverage": 0.0, "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": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "rank": null, "score": null}, "method": "attention-ranking-v13", "rank": null, "score": null}}, "readiness": "Paper only", "recognitionConfidence": 0.55, "relation": "evaluates_only", "releasedAt": "2026-08-05", "source": {"authors": ["Zhixiang Liang", "Yifei Liu", "Yidan Huang", "Haozhe Zhao", "Beichen Huang", "Jiaqi Wang", "Nan Duan", "Qiong Cao"], "categories": ["cs.AI"], "id": "2608.05212", "title": "SearchAuditor: Auditing and Attributing Failures in Long-Horizon Search Agents", "type": "arxiv", "url": "https://arxiv.org/abs/2608.05212"}, "sourceUpdatedAt": "2026-08-05T09:09:00Z", "topics": ["cs.AI"], "venueAttempts": [], "whyItMatters": "It provides a structured way to measure auditing quality for complex search-agent failures, reducing reliance on manual inspection of lengthy traces."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
