{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5DJPUUQQ4RMVG7SOW3CU5UB4LV","short_pith_number":"pith:5DJPUUQQ","schema_version":"1.0","canonical_sha256":"e8d2fa5210e459537e4eb6c54ed03c5d51b69511f293b780ad865a099ed45fea","source":{"kind":"arxiv","id":"2505.24105","version":1},"attestation_state":"computed","paper":{"title":"Training LLMs for EHR-Based Reasoning Tasks via Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jiacheng Lin, Jimeng Sun, Zhenbang Wu","submitted_at":"2025-05-30T01:13:22Z","abstract_excerpt":"We present EHRMIND, a practical recipe for adapting large language models (LLMs) to complex clinical reasoning tasks using reinforcement learning with verifiable rewards (RLVR). While RLVR has succeeded in mathematics and coding, its application to healthcare contexts presents unique challenges due to the specialized knowledge and reasoning required for electronic health record (EHR) interpretation. Our pilot study on the MEDCALC benchmark reveals two key failure modes: (1) misapplied knowledge, where models possess relevant medical knowledge but apply it incorrectly, and (2) missing knowledge"},"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":"2505.24105","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-30T01:13:22Z","cross_cats_sorted":[],"title_canon_sha256":"098d785e4fc5058fc44a284e4dd3b1f91680ce02a40e181bc85c227f076bd641","abstract_canon_sha256":"174f77e8f88f6ddca5bc5f299bc20c531ac288067fad26a643fb6bd0a6ee69f5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:12:41.949396Z","signature_b64":"E7SYeRuttqOgD0Yx2tgk1+Q6dh3uB6DMPVW6DTtoHm6gvfBefHkowKQLFi0kyiDpX1iUBH3mUFU/iiz0NHh8AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e8d2fa5210e459537e4eb6c54ed03c5d51b69511f293b780ad865a099ed45fea","last_reissued_at":"2026-07-05T11:12:41.948870Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:12:41.948870Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Training LLMs for EHR-Based Reasoning Tasks via Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jiacheng Lin, Jimeng Sun, Zhenbang Wu","submitted_at":"2025-05-30T01:13:22Z","abstract_excerpt":"We present EHRMIND, a practical recipe for adapting large language models (LLMs) to complex clinical reasoning tasks using reinforcement learning with verifiable rewards (RLVR). While RLVR has succeeded in mathematics and coding, its application to healthcare contexts presents unique challenges due to the specialized knowledge and reasoning required for electronic health record (EHR) interpretation. Our pilot study on the MEDCALC benchmark reveals two key failure modes: (1) misapplied knowledge, where models possess relevant medical knowledge but apply it incorrectly, and (2) missing knowledge"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.24105","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/2505.24105/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":"2505.24105","created_at":"2026-07-05T11:12:41.948934+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.24105v1","created_at":"2026-07-05T11:12:41.948934+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.24105","created_at":"2026-07-05T11:12:41.948934+00:00"},{"alias_kind":"pith_short_12","alias_value":"5DJPUUQQ4RMV","created_at":"2026-07-05T11:12:41.948934+00:00"},{"alias_kind":"pith_short_16","alias_value":"5DJPUUQQ4RMVG7SO","created_at":"2026-07-05T11:12:41.948934+00:00"},{"alias_kind":"pith_short_8","alias_value":"5DJPUUQQ","created_at":"2026-07-05T11:12:41.948934+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2512.19691","citing_title":"Scalable Stewardship of an LLM-Assisted Clinical Benchmark with Physician Oversight","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2603.23964","citing_title":"From Pixels to Digital Agents: An Empirical Study on the Taxonomy and Technological Trends of Reinforcement Learning Environments","ref_index":183,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5DJPUUQQ4RMVG7SOW3CU5UB4LV","json":"https://pith.science/pith/5DJPUUQQ4RMVG7SOW3CU5UB4LV.json","graph_json":"https://pith.science/api/pith-number/5DJPUUQQ4RMVG7SOW3CU5UB4LV/graph.json","events_json":"https://pith.science/api/pith-number/5DJPUUQQ4RMVG7SOW3CU5UB4LV/events.json","paper":"https://pith.science/paper/5DJPUUQQ"},"agent_actions":{"view_html":"https://pith.science/pith/5DJPUUQQ4RMVG7SOW3CU5UB4LV","download_json":"https://pith.science/pith/5DJPUUQQ4RMVG7SOW3CU5UB4LV.json","view_paper":"https://pith.science/paper/5DJPUUQQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.24105&json=true","fetch_graph":"https://pith.science/api/pith-number/5DJPUUQQ4RMVG7SOW3CU5UB4LV/graph.json","fetch_events":"https://pith.science/api/pith-number/5DJPUUQQ4RMVG7SOW3CU5UB4LV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5DJPUUQQ4RMVG7SOW3CU5UB4LV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5DJPUUQQ4RMVG7SOW3CU5UB4LV/action/storage_attestation","attest_author":"https://pith.science/pith/5DJPUUQQ4RMVG7SOW3CU5UB4LV/action/author_attestation","sign_citation":"https://pith.science/pith/5DJPUUQQ4RMVG7SOW3CU5UB4LV/action/citation_signature","submit_replication":"https://pith.science/pith/5DJPUUQQ4RMVG7SOW3CU5UB4LV/action/replication_record"}},"created_at":"2026-07-05T11:12:41.948934+00:00","updated_at":"2026-07-05T11:12:41.948934+00:00"}