{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2605.22219", "kind": "paper", "url": "https://arxiv.org/abs/2605.22219"}, {"id": "hf:pkuaiweb/sgr-bench", "kind": "dataset", "url": "https://huggingface.co/datasets/PKUAIWeb/SGR-BENCH"}], "authors": ["Ningyuan Li", "Haiyang Shen", "Mugeng Liu", "Yudong Han", "Zhuofan Shi", "Sixiong Xie", "Yun Ma"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2605.22219", "source_url": "https://arxiv.org/abs/2605.22219"}, "categories": ["General AI", "Information retrieval", "Language & Knowledge", "cs.AI"], "description": {"en": "SGR-Bench evaluates search agents on state-gated retrieval tasks. It includes 100 expert-curated tasks across 12 public data ecosystems, requiring agents to configure site-specific filters, views, hierarchies, or scopes to retrieve structured answers. Tasks come in goal-oriented and constraint-guided formulations."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2605.22219", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2605.22219", "title": "SGR-Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Beijing University of Technology", "Peking University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2605.22219", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2605.22219", "source_url": "https://openalex.org/W7162133976"}, "key": "claire-radar:2605.22219", "modality": null, "name": "SGR-Bench", "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_sgr-bench_d4ee6524", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "a55e28b79421355fcf4bce9653de07a79c05afc90bb7e437df75534eb2438698", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2605.22219"}, "publisher": null, "released": "2026-05-21", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2605.22219", "source_url": "https://arxiv.org/abs/2605.22219"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2605-22219", "source": "claire_radar", "source_benchmark_id": "2605.22219", "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": "PKUAIWeb/SGR-BENCH", "hfDatasetDownloads": 119, "hfDatasetLikes": 0, "hfPaperUpvotes": 0, "hfPaperUrl": "https://huggingface.co/papers/2605.22219", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetLikes": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfPaperUpvotes": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}}, "status": "fresh"}, "capabilities": ["Information retrieval"], "confidence": "High", "construction": "Transform Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "918576c8c087f4bc05fafe7baaf735ead5edcb0ab74b8f5ecfb8391d8c75d9a4", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "SGR-Bench evaluates search agents on state-gated retrieval tasks. It includes 100 expert-curated tasks across 12 public data ecosystems, requiring agents to configure site-specific filters, views, hierarchies, or scopes to retrieve structured answers. Tasks come in goal-oriented and constraint-guided formulations.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We introduce SGR-Bench, a benchmark for this setting containing 100 expert-curated tasks spanning six source families and 12 public data ecosystems."}, "familyId": "bmf_3984c874ff5c", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_sgr-bench_d4ee6524", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": "https://huggingface.co/datasets/PKUAIWeb/SGR-BENCH", "hfPaper": "https://huggingface.co/papers/2605.22219", "paper": "https://arxiv.org/abs/2605.22219", "pdf": "https://arxiv.org/pdf/2605.22219", "project": null, "report": "https://arxiv.org/abs/2605.22219"}, "metrics": [], "motivation": "Recent advances in large language models and tool-using agents have expanded the range of benchmarked web tasks.", "name": "SGR-Bench", "oneLine": "SGR-Bench evaluates search agents on state-gated retrieval tasks. It includes 100 expert-curated tasks across 12 public data ecosystems, requiring agents to configure site-specific filters, views, hierarchies, or scopes to retrieve structured answers. Tasks come in goal-oriented and constraint-guided formulations.", "paperTitle": "SGR-Bench: Benchmarking Search Agents on State-Gated Retrieval", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2605.22219", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "PKUAIWeb", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://huggingface.co/datasets/PKUAIWeb/SGR-BENCH"}], "ranking": {}, "readiness": "Inspectable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-05-21", "source": {"authors": ["Ningyuan Li", "Haiyang Shen", "Mugeng Liu", "Yudong Han", "Zhuofan Shi", "Sixiong Xie", "and Yun Ma"], "categories": ["cs.AI"], "id": "2605.22219", "title": "SGR-Bench: Benchmarking Search Agents on State-Gated Retrieval", "type": "arxiv", "url": "https://arxiv.org/abs/2605.22219"}, "sourceUpdatedAt": "2026-05-21T09:22:48Z", "topics": ["cs.AI"], "venueAttempts": [], "whyItMatters": "SGR-Bench addresses an undercharacterized class of retrieval tasks where evidence is hidden behind site-specific retrieval states. It provides a standardized evaluation to measure agents' ability to establish correct retrieval states, which is critical for real-world data retrieval from specialized websites."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
