{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CERNBCAMGHFXI5HJAXEXMFPU7L","short_pith_number":"pith:CERNBCAM","schema_version":"1.0","canonical_sha256":"1122d0880c31cb7474e905c97615f4fadd58f18a7c26ebd935687349e8b046bb","source":{"kind":"arxiv","id":"2509.04482","version":2},"attestation_state":"computed","paper":{"title":"Energy Landscapes Enable Reliable Abstention in Retrieval-Augmented Large Language Models for Healthcare","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Alex Silverthorne, Beth Albert, Gabriel Davis Jones, Lin Li, Magdalena Bachmann, Ravi Shankar, Sheng Wong","submitted_at":"2025-08-31T10:20:48Z","abstract_excerpt":"Reliable abstention is critical for retrieval-augmented generation (RAG) systems, particularly in safety-critical domains such as women's health, where incorrect answers can lead to harm. We present an energy-based model (EBM) that learns a smooth energy landscape over a dense semantic corpus of 2.6M guideline-derived questions, enabling the system to decide when to generate or abstain. We benchmark the EBM against a calibrated softmax baseline and a k-nearest neighbour (kNN) density heuristic across both easy and hard abstention splits, where hard cases are semantically challenging near-distr"},"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":"2509.04482","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-31T10:20:48Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9a0642c8cd330f0a440619b4ea2de7ccdd4373805c497f87ca05573e64512759","abstract_canon_sha256":"5dc3deaab580aa55b57ec635341dd4f320aec92dd6d553fd48a557b5b306e0d4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:06:49.734176Z","signature_b64":"U+otZJwgJdszYUGJT0KMaiVFR0LN0dUXY+WKfmvhiK6S056WzXhOuUkZjkVx75xbJiARgo35FrP9jUyoHm/KCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1122d0880c31cb7474e905c97615f4fadd58f18a7c26ebd935687349e8b046bb","last_reissued_at":"2026-07-05T12:06:49.733664Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:06:49.733664Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Energy Landscapes Enable Reliable Abstention in Retrieval-Augmented Large Language Models for Healthcare","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Alex Silverthorne, Beth Albert, Gabriel Davis Jones, Lin Li, Magdalena Bachmann, Ravi Shankar, Sheng Wong","submitted_at":"2025-08-31T10:20:48Z","abstract_excerpt":"Reliable abstention is critical for retrieval-augmented generation (RAG) systems, particularly in safety-critical domains such as women's health, where incorrect answers can lead to harm. We present an energy-based model (EBM) that learns a smooth energy landscape over a dense semantic corpus of 2.6M guideline-derived questions, enabling the system to decide when to generate or abstain. We benchmark the EBM against a calibrated softmax baseline and a k-nearest neighbour (kNN) density heuristic across both easy and hard abstention splits, where hard cases are semantically challenging near-distr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.04482","kind":"arxiv","version":2},"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/2509.04482/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":"2509.04482","created_at":"2026-07-05T12:06:49.733733+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.04482v2","created_at":"2026-07-05T12:06:49.733733+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.04482","created_at":"2026-07-05T12:06:49.733733+00:00"},{"alias_kind":"pith_short_12","alias_value":"CERNBCAMGHFX","created_at":"2026-07-05T12:06:49.733733+00:00"},{"alias_kind":"pith_short_16","alias_value":"CERNBCAMGHFXI5HJ","created_at":"2026-07-05T12:06:49.733733+00:00"},{"alias_kind":"pith_short_8","alias_value":"CERNBCAM","created_at":"2026-07-05T12:06:49.733733+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.17830","citing_title":"Remembering More, Risking More: Longitudinal Safety Risks in Memory-Equipped LLM Agents","ref_index":58,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CERNBCAMGHFXI5HJAXEXMFPU7L","json":"https://pith.science/pith/CERNBCAMGHFXI5HJAXEXMFPU7L.json","graph_json":"https://pith.science/api/pith-number/CERNBCAMGHFXI5HJAXEXMFPU7L/graph.json","events_json":"https://pith.science/api/pith-number/CERNBCAMGHFXI5HJAXEXMFPU7L/events.json","paper":"https://pith.science/paper/CERNBCAM"},"agent_actions":{"view_html":"https://pith.science/pith/CERNBCAMGHFXI5HJAXEXMFPU7L","download_json":"https://pith.science/pith/CERNBCAMGHFXI5HJAXEXMFPU7L.json","view_paper":"https://pith.science/paper/CERNBCAM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.04482&json=true","fetch_graph":"https://pith.science/api/pith-number/CERNBCAMGHFXI5HJAXEXMFPU7L/graph.json","fetch_events":"https://pith.science/api/pith-number/CERNBCAMGHFXI5HJAXEXMFPU7L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CERNBCAMGHFXI5HJAXEXMFPU7L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CERNBCAMGHFXI5HJAXEXMFPU7L/action/storage_attestation","attest_author":"https://pith.science/pith/CERNBCAMGHFXI5HJAXEXMFPU7L/action/author_attestation","sign_citation":"https://pith.science/pith/CERNBCAMGHFXI5HJAXEXMFPU7L/action/citation_signature","submit_replication":"https://pith.science/pith/CERNBCAMGHFXI5HJAXEXMFPU7L/action/replication_record"}},"created_at":"2026-07-05T12:06:49.733733+00:00","updated_at":"2026-07-05T12:06:49.733733+00:00"}