{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2607.16229", "kind": "paper", "url": "https://arxiv.org/abs/2607.16229"}], "authors": ["Rishab Ghosh", "Vinay Devarakonda"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2607.16229", "source_url": "https://arxiv.org/abs/2607.16229"}, "categories": ["Finance", "Safety & Trustworthiness", "stat.AP"], "description": {"en": "FinBench is a benchmark for evaluating calibration and uncertainty in financial forecasting with time-gated tasks, requiring probability of positive return and 80% prediction interval, scored with Brier and Winkler scores."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2607.16229", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2607.16229", "title": "FinBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["The University of Texas at Austin"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2607.16229", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2607.16229", "source_url": "https://openalex.org/W7169830300"}, "key": "claire-radar:2607.16229", "modality": null, "name": "FinBench", "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_finbench_34ca17f0", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "f6792d33711e852e89d3f2d293bdedb29c552a6f2b1a877d6c5464a91386d496", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2607.16229"}, "publisher": null, "released": "2026-06-24", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2607.16229", "source_url": "https://arxiv.org/abs/2607.16229"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2607-16229", "source": "claire_radar", "source_benchmark_id": "2607.16229", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["Finance"], "area": "Safety & Trustworthiness", "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/2607.16229", "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": "2f20475c33eaeac878e03ec82af798089271947533fa51c38bab9656d02aeecf", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "FinBench is a benchmark for evaluating calibration and uncertainty in financial forecasting with time-gated tasks, requiring probability of positive return and 80% prediction interval, scored with Brier and Winkler scores.", "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"], "snippet": "We introduce FinBench, a benchmark designed to evaluate calibration and uncertainty quality for financial forecasting in a setting that is (i) strictly time-gated to avoid look-ahead bias and (ii) evaluated with strictly proper scoring rules that penalize hallucinated confidence."}, "familyId": "bmf_5ee7524fb0ac", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_finbench_34ca17f0", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Financial Services"], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2607.16229", "paper": "https://arxiv.org/abs/2607.16229", "pdf": "https://arxiv.org/pdf/2607.16229", "project": null, "report": "https://arxiv.org/abs/2607.16229"}, "metrics": [], "motivation": "Large language models (LLMs) are increasingly used as components of agentic systems that observe, plan, and act.", "name": "FinBench", "oneLine": "FinBench is a benchmark for evaluating calibration and uncertainty in financial forecasting with time-gated tasks, requiring probability of positive return and 80% prediction interval, scored with Brier and Winkler scores.", "paperTitle": "FinBench: Time-Gated Calibration and Uncertainty Benchmarking for Agentic Financial Forecasting", "primaryDomain": "Finance", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2607.16229", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-06-24", "source": {"authors": ["Rishab Ghosh and Vinay Devarakonda"], "categories": ["stat.AP", "cs.LG", "q-fin.CP", "q-fin.PM"], "id": "2607.16229", "title": "FinBench: Time-Gated Calibration and Uncertainty Benchmarking for Agentic Financial Forecasting", "type": "arxiv", "url": "https://arxiv.org/abs/2607.16229"}, "sourceUpdatedAt": "2026-06-24T01:27:30Z", "topics": ["stat.AP"], "venueAttempts": [], "whyItMatters": "Financial forecasting agents risk overconfidence; FinBench addresses this by testing probabilistic calibration under temporal constraints, but its pilot scale limits immediate comparison value."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
