{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2607.11079", "kind": "paper", "url": "https://arxiv.org/abs/2607.11079"}], "authors": ["Chuhan Shi", "Xiaoquan Ren", "Sicheng Song", "Haobo Li", "Rui Sheng", "Yushi Sun"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2607.11079", "source_url": "https://arxiv.org/abs/2607.11079"}, "categories": ["AI Scientist", "General AI", "Language & Knowledge", "Materials & Chemistry"], "description": {"en": "SDABench evaluates LLMs' scientific data analysis capabilities across six capabilities (descriptive, exploratory, inferential, predictive, causal, mechanistic) and five domains, with 527 real and 6000 synthetic instances in multiple-choice and open-ended formats."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2607.11079", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2607.11079", "title": "SDABench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Southeast University", "East China Normal University", "Hong Kong University of Science and Technology"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2607.11079", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2607.11079", "source_url": "https://openalex.org/W7168272912"}, "key": "claire-radar:2607.11079", "modality": null, "name": "SDABench", "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_sdabench_9285d4fe", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "9390bb6d77e9d6ab270abfe6df713f11bf788ef3cc9caa62d5b2dfbd7e4c20bd", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2607.11079"}, "publisher": null, "released": "2026-07-13", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2607.11079", "source_url": "https://arxiv.org/abs/2607.11079"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2607-11079", "source": "claire_radar", "source_benchmark_id": "2607.11079", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Machine Generated", "applicationDomains": ["General AI", "Materials & Chemistry"], "area": "Language & Knowledge", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": "2026-07-15T00:00:00.000Z", "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 4, "hfPaperUrl": "https://huggingface.co/papers/2607.11079", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}}, "status": "fresh"}, "capabilities": [], "confidence": "High", "construction": "Original Synthetic", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "07fdbbabecc9cf01fcdb9fc894ad3d1ad4a37be1b3a46ad553feef967c6d090b", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "SDABench evaluates LLMs' scientific data analysis capabilities across six capabilities (descriptive, exploratory, inferential, predictive, causal, mechanistic) and five domains, with 527 real and 6000 synthetic instances in multiple-choice and open-ended formats.", "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": "We introduce SDABench, a benchmark that reorganizes evaluation around six capabilities (descriptive, exploratory, inferential, predictive, causal, and mechanistic) across five domains (Biology, Chemistry, Environment, Geography, Physics)."}, "familyId": "bmf_4b6f310c64ce", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_sdabench_9285d4fe", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Materials & Chemicals"], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2607.11079", "paper": "https://arxiv.org/abs/2607.11079", "pdf": "https://arxiv.org/pdf/2607.11079", "project": null, "report": "https://arxiv.org/abs/2607.11079"}, "metrics": [], "motivation": "Existing benchmarks for scientific data analysis evaluate LLMs primarily on code execution or workflow completion, overlooking that scientific analysis serves to support distinct types of scientific claims: hypothesis exploration, statistical inference, mechanistic explanation, each with different assumptions and validity criteria.", "name": "SDABench", "oneLine": "SDABench evaluates LLMs' scientific data analysis capabilities across six capabilities (descriptive, exploratory, inferential, predictive, causal, mechanistic) and five domains, with 527 real and 6000 synthetic instances in multiple-choice and open-ended formats.", "paperTitle": "Are LLMs Ready for Scientific Discovery? A Capability-Oriented Benchmark for AI Scientists", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2607.11079", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {"90d": {"confidence": "Low", "coverage": 0.15, "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": 0.2843790034276272, "value": 4}}, "confidence": "Low", "coverage": 0.15, "rank": null, "score": 47}, "method": "attention-ranking-v13", "rank": null, "score": 47}}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-07-13", "source": {"authors": ["Chuhan Shi", "Xiaoquan Ren", "Sicheng Song", "Haobo Li", "Rui Sheng", "Yushi Sun"], "categories": ["cs.AI"], "id": "2607.11079", "title": "Are LLMs Ready for Scientific Discovery? A Capability-Oriented Benchmark for AI Scientists", "type": "arxiv", "url": "https://arxiv.org/abs/2607.11079"}, "sourceUpdatedAt": "2026-07-13T04:39:40Z", "topics": ["AI Scientist"], "venueAttempts": [], "whyItMatters": "This benchmark reveals that LLMs degrade sharply on tasks requiring assumption selection, latent-process modeling, and mechanistic reasoning, highlighting gaps for scientific discovery applications."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
