{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5J3NGKAFW7H3DUSI6U4XUNDJFL","short_pith_number":"pith:5J3NGKAF","schema_version":"1.0","canonical_sha256":"ea76d32805b7cfb1d248f5397a34692afd0998d82561fd50f9ca2007fdd20a52","source":{"kind":"arxiv","id":"2504.18080","version":1},"attestation_state":"computed","paper":{"title":"Stabilizing Reasoning in Medical LLMs with Continued Pretraining and Reasoning Preference Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Junichiro Iwasawa, Keita Suzuki, Wataru Kawakami","submitted_at":"2025-04-25T05:15:31Z","abstract_excerpt":"Large Language Models (LLMs) show potential in medicine, yet clinical adoption is hindered by concerns over factual accuracy, language-specific limitations (e.g., Japanese), and critically, their reliability when required to generate reasoning explanations -- a prerequisite for trust. This paper introduces Preferred-MedLLM-Qwen-72B, a 72B-parameter model optimized for the Japanese medical domain to achieve both high accuracy and stable reasoning. We employ a two-stage fine-tuning process on the Qwen2.5-72B base model: first, Continued Pretraining (CPT) on a comprehensive Japanese medical corpu"},"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":"2504.18080","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-04-25T05:15:31Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"cddf5cbfe4ac02c20fb77d8f3645fbd6bd57961e61f517c834ca585c67cadfee","abstract_canon_sha256":"e509aa12e9f6bbd3f6804f4007b2a99a5e7da8133b8717d6a0bbc6d6ebb3561c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:54:01.281726Z","signature_b64":"/dOptZ92dHwd0j/vc8HsH96cpiUA5wnsspWbu5L23bVD8W3cxwQG47Q/vGmUIadK5PuNGC4b/9dPMoSC9UGsBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ea76d32805b7cfb1d248f5397a34692afd0998d82561fd50f9ca2007fdd20a52","last_reissued_at":"2026-07-05T10:54:01.281189Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:54:01.281189Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Stabilizing Reasoning in Medical LLMs with Continued Pretraining and Reasoning Preference Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Junichiro Iwasawa, Keita Suzuki, Wataru Kawakami","submitted_at":"2025-04-25T05:15:31Z","abstract_excerpt":"Large Language Models (LLMs) show potential in medicine, yet clinical adoption is hindered by concerns over factual accuracy, language-specific limitations (e.g., Japanese), and critically, their reliability when required to generate reasoning explanations -- a prerequisite for trust. This paper introduces Preferred-MedLLM-Qwen-72B, a 72B-parameter model optimized for the Japanese medical domain to achieve both high accuracy and stable reasoning. We employ a two-stage fine-tuning process on the Qwen2.5-72B base model: first, Continued Pretraining (CPT) on a comprehensive Japanese medical corpu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.18080","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/2504.18080/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":"2504.18080","created_at":"2026-07-05T10:54:01.281244+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.18080v1","created_at":"2026-07-05T10:54:01.281244+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.18080","created_at":"2026-07-05T10:54:01.281244+00:00"},{"alias_kind":"pith_short_12","alias_value":"5J3NGKAFW7H3","created_at":"2026-07-05T10:54:01.281244+00:00"},{"alias_kind":"pith_short_16","alias_value":"5J3NGKAFW7H3DUSI","created_at":"2026-07-05T10:54:01.281244+00:00"},{"alias_kind":"pith_short_8","alias_value":"5J3NGKAF","created_at":"2026-07-05T10:54:01.281244+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.09584","citing_title":"CLR-voyance: Reinforcing Open-Ended Reasoning for Inpatient Clinical Decision Support with Outcome-Aware Rubrics","ref_index":32,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5J3NGKAFW7H3DUSI6U4XUNDJFL","json":"https://pith.science/pith/5J3NGKAFW7H3DUSI6U4XUNDJFL.json","graph_json":"https://pith.science/api/pith-number/5J3NGKAFW7H3DUSI6U4XUNDJFL/graph.json","events_json":"https://pith.science/api/pith-number/5J3NGKAFW7H3DUSI6U4XUNDJFL/events.json","paper":"https://pith.science/paper/5J3NGKAF"},"agent_actions":{"view_html":"https://pith.science/pith/5J3NGKAFW7H3DUSI6U4XUNDJFL","download_json":"https://pith.science/pith/5J3NGKAFW7H3DUSI6U4XUNDJFL.json","view_paper":"https://pith.science/paper/5J3NGKAF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.18080&json=true","fetch_graph":"https://pith.science/api/pith-number/5J3NGKAFW7H3DUSI6U4XUNDJFL/graph.json","fetch_events":"https://pith.science/api/pith-number/5J3NGKAFW7H3DUSI6U4XUNDJFL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5J3NGKAFW7H3DUSI6U4XUNDJFL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5J3NGKAFW7H3DUSI6U4XUNDJFL/action/storage_attestation","attest_author":"https://pith.science/pith/5J3NGKAFW7H3DUSI6U4XUNDJFL/action/author_attestation","sign_citation":"https://pith.science/pith/5J3NGKAFW7H3DUSI6U4XUNDJFL/action/citation_signature","submit_replication":"https://pith.science/pith/5J3NGKAFW7H3DUSI6U4XUNDJFL/action/replication_record"}},"created_at":"2026-07-05T10:54:01.281244+00:00","updated_at":"2026-07-05T10:54:01.281244+00:00"}