{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.11447", "kind": "paper", "url": "https://arxiv.org/abs/2606.11447"}], "authors": ["Meysam Alizadeh", "Mohsen Mosleh", "Fabrizio Gilardi", "Atoosa Kasirzadeh", "Joshua Tucker"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.11447", "source_url": "https://arxiv.org/abs/2606.11447"}, "categories": ["General AI", "Language & Knowledge", "cs.CL"], "description": {"en": "SocSci-Repro-Bench evaluates AI coding agents on reproducing social science findings from 221 tasks across four disciplines and 13 domains, using studies with known reproducibility outcomes."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.11447", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.11447", "title": "SocSci-Repro-Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["University of Oxford", "University of Zurich", "Carnegie Mellon University", "New York University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.11447", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.11447", "source_url": "https://openalex.org/W7164404586"}, "key": "claire-radar:2606.11447", "modality": null, "name": "SocSci-Repro-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_socsci-repro-bench_8a1a65b5", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "a23a4150ea7090dad91a162933f37b0d2c51fbe421d218460f3e4ba15c418500", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2606.11447"}, "publisher": null, "released": "2026-06-09", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.11447", "source_url": "https://arxiv.org/abs/2606.11447"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-11447", "source": "claire_radar", "source_benchmark_id": "2606.11447", "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/2606.11447", "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": "cead3849ff644a663b207bd72d807853980df1f2c8c6f7f62f7d4e77ab5f3a31", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "SocSci-Repro-Bench evaluates AI coding agents on reproducing social science findings from 221 tasks across four disciplines and 13 domains, using studies with known reproducibility outcomes.", "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": "Here we introduce SocSci-Repro-Bench, a benchmark of 221 tasks spanning four disciplines and 13 substantive domains, constructed from studies whose results are either fully reproducible with available materials or demonstrably non-reproducible due to missing data, allowing us to isolate agents' reproduction capacity."}, "familyId": "bmf_3424da46c0af", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_socsci-repro-bench_8a1a65b5", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2606.11447", "paper": "https://arxiv.org/abs/2606.11447", "pdf": "https://arxiv.org/pdf/2606.11447", "project": null, "report": "https://arxiv.org/abs/2606.11447"}, "metrics": [], "motivation": "Recent anecdotal evidence suggests that AI coding agents can reproduce published findings when provided with original data and code; yet systematic evaluation across social sciences remains limited.", "name": "SocSci-Repro-Bench", "oneLine": "SocSci-Repro-Bench evaluates AI coding agents on reproducing social science findings from 221 tasks across four disciplines and 13 domains, using studies with known reproducibility outcomes.", "paperTitle": "AI Coding Agents Can Reproduce Social Science Findings", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.11447", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "arXiv", "organizationType": "benchmark-organization", "role": "benchmark-publisher", "sourceUrl": "https://arxiv.org/abs/2606.11447"}], "ranking": {}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-06-09", "source": {"authors": ["Meysam Alizadeh", "Mohsen Mosleh", "Fabrizio Gilardi", "Atoosa Kasirzadeh", "Joshua Tucker"], "categories": ["cs.CL"], "id": "2606.11447", "title": "AI Coding Agents Can Reproduce Social Science Findings", "type": "arxiv", "url": "https://arxiv.org/abs/2606.11447"}, "sourceUpdatedAt": "2026-06-09T21:00:05Z", "topics": ["cs.CL"], "venueAttempts": [], "whyItMatters": "SocSci-Repro-Bench addresses the lack of systematic evaluation of AI agents' ability to execute computational workflows, providing a protocol to isolate agent performance from issues in reproduction materials and informing practical use of agents in scientific production."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
