{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.25071", "kind": "paper", "url": "https://arxiv.org/abs/2608.25071"}, {"id": "gh:mindbench-ai/healthbench-psych", "kind": "repo", "url": "https://github.com/mindbench-ai/healthbench-psych"}, {"id": "hf:mindbench-ai/healthbench-psych", "kind": "dataset", "url": "https://huggingface.co/datasets/mindbench-ai/healthbench-psych"}], "authors": ["Matthew Flathers", "Phuong Anh Nguyen", "Jill Noorily", "Julian Herpertz", "Meiting Chen", "Jasreen Multani", "Samuel Powell", "Mason Granof", "Mark Kalinch", "John Torous"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.25071", "source_url": "https://arxiv.org/abs/2608.25071"}, "categories": ["Biology & Drug Discovery", "Language & Knowledge", "cs.CL"], "description": {"en": "Evaluates LLMs on 610 mental-health conversations from HealthBench using a three-judge panel and physician rubrics, with a harder 119-conversation subset."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.25071", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.25071", "title": "HealthBench-Psych"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Beth Israel Deaconess Medical Center"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2608.25071", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2608.25071", "source_url": "https://openalex.org/W7204449673"}, "key": "claire-radar:2608.25071", "modality": null, "name": "HealthBench-Psych", "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_healthbench-psych_83387771", "origin_source": "arxiv", "recognition_confidence": "0.6", "record_sha256": "487640921309b2296ea05d99075a0f17bf57f536b5317cc9370bf4966bd12605", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-08-27T04:14:52.836173Z", "source_url": "https://arxiv.org/abs/2608.25071"}, "publisher": null, "released": "2026-08-25", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.25071", "source_url": "https://arxiv.org/abs/2608.25071"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-25071", "source": "claire_radar", "source_benchmark_id": "2608.25071", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["Biology & Drug Discovery"], "area": "Language & Knowledge", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/mindbench-ai/healthbench-psych", "githubScope": "benchmark_repo", "githubStars": 2, "hfDailySubmittedAt": null, "hfDataset": "mindbench-ai/healthbench-psych", "hfDatasetDownloads": 367, "hfDatasetLikes": 2, "hfPaperUpvotes": 0, "hfPaperUrl": "https://huggingface.co/papers/2608.25071", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetDownloads": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-09-25", "state": "stale"}, "hfDatasetLikes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-09-25", "state": "stale"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-09-25", "state": "stale"}}, "status": "fresh"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "Mental-health LLM evaluation is a high-interest niche, and the released artifacts and evidence-backed rigor may drive moderate early attention.", "score": 55}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-27T04:14:52.836173Z", "inputHash": "9555af4373b5ee095401d81578f3d1ec04604f7487c9c76347e2ce2200884efb", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "HealthBench-Psych: A Mental Health Subset of OpenAI's HealthBench", "canonicalNameSource": "paper_title", "decisionReason": "Includes released subset, pipeline, model responses, grades, and analysis code; offers a stable scoring protocol and public reuse path despite requiring access to the original HealthBench corpus.", "model": "deepseek-v4-pro", "reviewedAt": "2026-08-27T04:14:52.836173Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates LLMs on 610 mental-health conversations from HealthBench using a three-judge panel and physician rubrics, with a harder 119-conversation subset.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["exact named benchmark artifact released in abstract", "public artifact URL"], "snippet": "We introduce HealthBench-Psych and HealthBench-Psych-Hard."}, "familyId": "bmf_d81079d8455c", "firstSeenAt": "2026-08-27", "heat": null, "id": "bm_healthbench-psych_83387771", "indexedAt": "2026-08-27T04:12:10.575570Z", "industrySectors": ["Pharma & Biotech"], "links": {"code": "https://github.com/mindbench-ai/healthbench-psych", "data": "https://huggingface.co/datasets/mindbench-ai/healthbench-psych", "paper": "https://arxiv.org/abs/2608.25071", "pdf": "https://arxiv.org/pdf/2608.25071", "project": null, "report": "https://arxiv.org/abs/2608.25071"}, "metrics": [], "motivation": "General-purpose health benchmarks increasingly anchor claims about LLM medical performance, but they are not always resolved by clinical specialty, making domain-specific performance hard to isolate.", "name": "HealthBench-Psych", "oneLine": "Evaluates LLMs on 610 mental-health conversations from HealthBench using a three-judge panel and physician rubrics, with a harder 119-conversation subset.", "paperTitle": "HealthBench-Psych: A Mental Health Subset of OpenAI's HealthBench", "primaryDomain": "Biology & Drug Discovery", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.25071", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-27T04:12:10.575570Z"}, "publications": [], "ranking": {"90d": {"confidence": "High", "coverage": 1.0, "datasetDownloadRank": 32, "datasetRankPopulation": 85, "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": 0.1466449917496476, "value": 2}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": 0.6084139253616965, "value": 367}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.0, "value": 0}}, "confidence": "High", "coverage": 1.0, "rank": 360, "score": 26}, "method": "attention-ranking-v13", "rank": 360, "score": 26}}, "readiness": "Runnable", "recognitionConfidence": 0.6, "relation": "introduces", "releasedAt": "2026-08-25", "source": {"authors": ["Matthew Flathers", "Phuong Anh Nguyen", "Jill Noorily", "Julian Herpertz", "Meiting Chen", "Jasreen Multani", "Samuel Powell", "Mason Granof", "Mark Kalinch", "John Torous"], "categories": ["cs.CL", "cs.AI", "cs.CY"], "id": "2608.25071", "title": "HealthBench-Psych: A Mental Health Subset of OpenAI's HealthBench", "type": "arxiv", "url": "https://arxiv.org/abs/2608.25071"}, "sourceUpdatedAt": "2026-08-25T19:05:19Z", "topics": ["cs.CL"], "venueAttempts": [], "whyItMatters": "Provides a clinically validated mental-health evaluation subset with released data, code, and judge panel, enabling reproducible specialty-specific model assessment."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
