{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:3YYXKJBMBEOBPF6MWCYH6LPY3Y","short_pith_number":"pith:3YYXKJBM","schema_version":"1.0","canonical_sha256":"de3175242c091c1797ccb0b07f2df8de162f70a50a0b4fb1f915aa1dfba64e10","source":{"kind":"arxiv","id":"2606.22419","version":1},"attestation_state":"computed","paper":{"title":"Knowledge-Graph Grounding Helps LLMs Only for Out-of-Training Knowledge: A Controlled Study on Clinical Question Answering","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DB"],"primary_cat":"cs.CL","authors_text":"Madhulatha Mandarapu, Sandeep Kunkunuru","submitted_at":"2026-06-21T10:12:17Z","abstract_excerpt":"A recent Nature Medicine study reports that general-purpose frontier LLMs outperform specialized retrieval-augmented clinical tools on medical benchmarks, and that retrieval can hurt strong models. We ask the natural follow-up: does structured knowledge-graph (KG) grounding change this, and when does grounding help at all? We contribute two results. First, a reproduction: the study's headline HealthBench score (~88) is the Consensus variant, not full HealthBench, where frontier models and ideal completions both score ~46-47 under a physician-calibrated grader (agreement 82.5%); we reproduce GP"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.22419","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-21T10:12:17Z","cross_cats_sorted":["cs.DB"],"title_canon_sha256":"06ccf458ac6b0c3924caca37cf7e68f0ab28ddd46c054e4831bac904e2fb46d2","abstract_canon_sha256":"08bc8776ae341b8415879f372aa8671d80e1a8f4918af3e32f24135571c100ac"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T02:13:37.935673Z","signature_b64":"yowzBmdYsoY1i8GlK54vzfEvFP/NvOa/snbAqEXVQ3L8kXz+W6nNSOicuctIBD4Dxg60KTQVWKAPSJIsTw9iCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"de3175242c091c1797ccb0b07f2df8de162f70a50a0b4fb1f915aa1dfba64e10","last_reissued_at":"2026-06-23T02:13:37.935180Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T02:13:37.935180Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Knowledge-Graph Grounding Helps LLMs Only for Out-of-Training Knowledge: A Controlled Study on Clinical Question Answering","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DB"],"primary_cat":"cs.CL","authors_text":"Madhulatha Mandarapu, Sandeep Kunkunuru","submitted_at":"2026-06-21T10:12:17Z","abstract_excerpt":"A recent Nature Medicine study reports that general-purpose frontier LLMs outperform specialized retrieval-augmented clinical tools on medical benchmarks, and that retrieval can hurt strong models. We ask the natural follow-up: does structured knowledge-graph (KG) grounding change this, and when does grounding help at all? We contribute two results. First, a reproduction: the study's headline HealthBench score (~88) is the Consensus variant, not full HealthBench, where frontier models and ideal completions both score ~46-47 under a physician-calibrated grader (agreement 82.5%); we reproduce GP"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.22419","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2606.22419/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.22419","created_at":"2026-06-23T02:13:37.935247+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.22419v1","created_at":"2026-06-23T02:13:37.935247+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.22419","created_at":"2026-06-23T02:13:37.935247+00:00"},{"alias_kind":"pith_short_12","alias_value":"3YYXKJBMBEOB","created_at":"2026-06-23T02:13:37.935247+00:00"},{"alias_kind":"pith_short_16","alias_value":"3YYXKJBMBEOBPF6M","created_at":"2026-06-23T02:13:37.935247+00:00"},{"alias_kind":"pith_short_8","alias_value":"3YYXKJBM","created_at":"2026-06-23T02:13:37.935247+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3YYXKJBMBEOBPF6MWCYH6LPY3Y","json":"https://pith.science/pith/3YYXKJBMBEOBPF6MWCYH6LPY3Y.json","graph_json":"https://pith.science/api/pith-number/3YYXKJBMBEOBPF6MWCYH6LPY3Y/graph.json","events_json":"https://pith.science/api/pith-number/3YYXKJBMBEOBPF6MWCYH6LPY3Y/events.json","paper":"https://pith.science/paper/3YYXKJBM"},"agent_actions":{"view_html":"https://pith.science/pith/3YYXKJBMBEOBPF6MWCYH6LPY3Y","download_json":"https://pith.science/pith/3YYXKJBMBEOBPF6MWCYH6LPY3Y.json","view_paper":"https://pith.science/paper/3YYXKJBM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.22419&json=true","fetch_graph":"https://pith.science/api/pith-number/3YYXKJBMBEOBPF6MWCYH6LPY3Y/graph.json","fetch_events":"https://pith.science/api/pith-number/3YYXKJBMBEOBPF6MWCYH6LPY3Y/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3YYXKJBMBEOBPF6MWCYH6LPY3Y/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3YYXKJBMBEOBPF6MWCYH6LPY3Y/action/storage_attestation","attest_author":"https://pith.science/pith/3YYXKJBMBEOBPF6MWCYH6LPY3Y/action/author_attestation","sign_citation":"https://pith.science/pith/3YYXKJBMBEOBPF6MWCYH6LPY3Y/action/citation_signature","submit_replication":"https://pith.science/pith/3YYXKJBMBEOBPF6MWCYH6LPY3Y/action/replication_record"}},"created_at":"2026-06-23T02:13:37.935247+00:00","updated_at":"2026-06-23T02:13:37.935247+00:00"}