{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QAJH2MII72FDAZMHP3X7EQUYF3","short_pith_number":"pith:QAJH2MII","schema_version":"1.0","canonical_sha256":"80127d3108fe8a3065877eeff242982ed59fdbbf6572700f6554db64adcb0aeb","source":{"kind":"arxiv","id":"2607.14385","version":1},"attestation_state":"computed","paper":{"title":"MamaBench: Benchmarking LLM Robustness in Maternal and Child Health Diagnosis through Counterfactual Clinical Perturbation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Abiodun Adereni, Angel Ezendu, Anuoluwa Sotome, Ifeoma Nwabueze, Oluwadarasimi Oguntuase, Oluwafunke Akinbuwa, Oluwasegun Oguntuase, Oluwaseun Odunsi, Samuel Okoko, Thanni Adewuyi","submitted_at":"2026-07-15T21:56:19Z","abstract_excerpt":"Large language models achieve strong scores on medical benchmarks, yet these benchmarks evaluate each question in isolation, providing no measure of whether a system can distinguish clinically similar presentations requiring different interventions. We introduce MamaBench, the first counterfactual benchmark for maternal and paediatric AI: 434 expert-authored clinical narratives in 217 pairs across 371 pathologies, evaluated via the Bias Trap Rate (BTR), the conditional probability that a model fails the counterfactual given success on the base case. We propose Evidence-Anchored RAG (EA-RAG), a"},"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":"2607.14385","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-15T21:56:19Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"792da533680eabc6ce9103b0dbc5f5f54a5b3a5af5bfc4bf6ee8d79acfb6002e","abstract_canon_sha256":"c51441026965d32fd624f5028d2393bbaf0f1693ec02f0e5256c46d8ac55543a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-17T00:21:10.032763Z","signature_b64":"LNsLz7sCUjuaGMBVVcZewqRKBE50zzFVCNGRlCNi+nJniHcbIrljvl5aBxkVeBRCJ2+1U8sTEIX8EULRBeTYBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"80127d3108fe8a3065877eeff242982ed59fdbbf6572700f6554db64adcb0aeb","last_reissued_at":"2026-07-17T00:21:10.031899Z","signature_status":"signed_v1","first_computed_at":"2026-07-17T00:21:10.031899Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MamaBench: Benchmarking LLM Robustness in Maternal and Child Health Diagnosis through Counterfactual Clinical Perturbation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Abiodun Adereni, Angel Ezendu, Anuoluwa Sotome, Ifeoma Nwabueze, Oluwadarasimi Oguntuase, Oluwafunke Akinbuwa, Oluwasegun Oguntuase, Oluwaseun Odunsi, Samuel Okoko, Thanni Adewuyi","submitted_at":"2026-07-15T21:56:19Z","abstract_excerpt":"Large language models achieve strong scores on medical benchmarks, yet these benchmarks evaluate each question in isolation, providing no measure of whether a system can distinguish clinically similar presentations requiring different interventions. We introduce MamaBench, the first counterfactual benchmark for maternal and paediatric AI: 434 expert-authored clinical narratives in 217 pairs across 371 pathologies, evaluated via the Bias Trap Rate (BTR), the conditional probability that a model fails the counterfactual given success on the base case. We propose Evidence-Anchored RAG (EA-RAG), a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.14385","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/2607.14385/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":"2607.14385","created_at":"2026-07-17T00:21:10.032336+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.14385v1","created_at":"2026-07-17T00:21:10.032336+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.14385","created_at":"2026-07-17T00:21:10.032336+00:00"},{"alias_kind":"pith_short_12","alias_value":"QAJH2MII72FD","created_at":"2026-07-17T00:21:10.032336+00:00"},{"alias_kind":"pith_short_16","alias_value":"QAJH2MII72FDAZMH","created_at":"2026-07-17T00:21:10.032336+00:00"},{"alias_kind":"pith_short_8","alias_value":"QAJH2MII","created_at":"2026-07-17T00:21:10.032336+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/QAJH2MII72FDAZMHP3X7EQUYF3","json":"https://pith.science/pith/QAJH2MII72FDAZMHP3X7EQUYF3.json","graph_json":"https://pith.science/api/pith-number/QAJH2MII72FDAZMHP3X7EQUYF3/graph.json","events_json":"https://pith.science/api/pith-number/QAJH2MII72FDAZMHP3X7EQUYF3/events.json","paper":"https://pith.science/paper/QAJH2MII"},"agent_actions":{"view_html":"https://pith.science/pith/QAJH2MII72FDAZMHP3X7EQUYF3","download_json":"https://pith.science/pith/QAJH2MII72FDAZMHP3X7EQUYF3.json","view_paper":"https://pith.science/paper/QAJH2MII","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.14385&json=true","fetch_graph":"https://pith.science/api/pith-number/QAJH2MII72FDAZMHP3X7EQUYF3/graph.json","fetch_events":"https://pith.science/api/pith-number/QAJH2MII72FDAZMHP3X7EQUYF3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QAJH2MII72FDAZMHP3X7EQUYF3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QAJH2MII72FDAZMHP3X7EQUYF3/action/storage_attestation","attest_author":"https://pith.science/pith/QAJH2MII72FDAZMHP3X7EQUYF3/action/author_attestation","sign_citation":"https://pith.science/pith/QAJH2MII72FDAZMHP3X7EQUYF3/action/citation_signature","submit_replication":"https://pith.science/pith/QAJH2MII72FDAZMHP3X7EQUYF3/action/replication_record"}},"created_at":"2026-07-17T00:21:10.032336+00:00","updated_at":"2026-07-17T00:21:10.032336+00:00"}