{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:IT2GD6TBP23FJ7VBXU7JLEFEVZ","short_pith_number":"pith:IT2GD6TB","canonical_record":{"source":{"id":"2607.09322","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-10T12:04:31Z","cross_cats_sorted":[],"title_canon_sha256":"648c9f0bf81196fc583ad4f319bf7de1ee1df0ff480596ec2fe802f6ae9cab93","abstract_canon_sha256":"d09fc8c21821920d28a77aea3098b1d168552b2ffc520bf6c6b35160d401b76c"},"schema_version":"1.0"},"canonical_sha256":"44f461fa617eb654fea1bd3e9590a4ae41d437cc21cc9413af6e68fd773c1bc1","source":{"kind":"arxiv","id":"2607.09322","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.09322","created_at":"2026-07-13T01:19:02Z"},{"alias_kind":"arxiv_version","alias_value":"2607.09322v1","created_at":"2026-07-13T01:19:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.09322","created_at":"2026-07-13T01:19:02Z"},{"alias_kind":"pith_short_12","alias_value":"IT2GD6TBP23F","created_at":"2026-07-13T01:19:02Z"},{"alias_kind":"pith_short_16","alias_value":"IT2GD6TBP23FJ7VB","created_at":"2026-07-13T01:19:02Z"},{"alias_kind":"pith_short_8","alias_value":"IT2GD6TB","created_at":"2026-07-13T01:19:02Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:IT2GD6TBP23FJ7VBXU7JLEFEVZ","target":"record","payload":{"canonical_record":{"source":{"id":"2607.09322","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-10T12:04:31Z","cross_cats_sorted":[],"title_canon_sha256":"648c9f0bf81196fc583ad4f319bf7de1ee1df0ff480596ec2fe802f6ae9cab93","abstract_canon_sha256":"d09fc8c21821920d28a77aea3098b1d168552b2ffc520bf6c6b35160d401b76c"},"schema_version":"1.0"},"canonical_sha256":"44f461fa617eb654fea1bd3e9590a4ae41d437cc21cc9413af6e68fd773c1bc1","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-13T01:19:02.339414Z","signature_b64":"lss3Ibn2J2N9snwE9Pp+ohaLrPNDUCtOvixXwvQGIUaM0qiDYICKfJi4arC1NP1suxPQKMmeQWyOPKSCvCHVBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"44f461fa617eb654fea1bd3e9590a4ae41d437cc21cc9413af6e68fd773c1bc1","last_reissued_at":"2026-07-13T01:19:02.338243Z","signature_status":"signed_v1","first_computed_at":"2026-07-13T01:19:02.338243Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2607.09322","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-13T01:19:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Q0HiOmBMasY93ARngzj8sOxCUVFOQBOEPPTz1YvojEPNxfHREofciMhK7aRC/vhEpgt0v1W7csQvMQ2XK5FgDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T17:12:31.886856Z"},"content_sha256":"b27a9b8149a8966517ae61f6c2a73c0bb7c3a674699b348d365330d9145931ac","schema_version":"1.0","event_id":"sha256:b27a9b8149a8966517ae61f6c2a73c0bb7c3a674699b348d365330d9145931ac"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:IT2GD6TBP23FJ7VBXU7JLEFEVZ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"LongMedBench: Benchmarking Medical Agents for Long-Horizon Clinical Decision-Making","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Hongxia Xu, Weiqi Zhai, Xiaocheng Zhang, Yanzhen Chen, Zhiting Fan, Zihan Xu, Zuozhu Liu","submitted_at":"2026-07-10T12:04:31Z","abstract_excerpt":"In this work, we introduce LongMedBench, a real-world EHR-based benchmark for long-horizon clinical decision-making. Prior evaluations of LLM-based medical agents have largely emphasized short-context knowledge QA and tool use. However, real-world medical care is inherently longitudinal, and clinicians must aggregate evidence across repeated visits, tests, and evolving treatments. Therefore, long-horizon interaction is essential for realistic assessment. LongMedBench is constructed via a reproducible pipeline that integrates MIMIC-IV admission records and clinical notes into time-series event "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.09322","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/2607.09322/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-13T01:19:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"fKH3CKi5vlOIxqxSoY03Ep+5CFPf8A+OVtrCn/mtmUe9rFt9LvY8zT7NoHysFfQR5ImhjN1PWju0B7c3gkX1Aw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T17:12:31.887457Z"},"content_sha256":"4005db8e63aaaade59c167d5a3dc946668e60517da231e73d164a301f546f956","schema_version":"1.0","event_id":"sha256:4005db8e63aaaade59c167d5a3dc946668e60517da231e73d164a301f546f956"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:IT2GD6TBP23FJ7VBXU7JLEFEVZ","target":"integrity","payload":{"note":"DOI '10.2196/84120' as printed in the bibliography is syntactically invalid and cannot resolve.","snippet":"Gong, E.J., Bang, C.S., Lee, J.J., Baik, G.H.: Knowledge-practice performance gap in clinical large language models: Systematic review of 39 benchmarks. Journal of medical Internet research 27, e84120 (2025). https://doi.org/10.2196/84120","arxiv_id":"2607.09322","detector":"doi_compliance","evidence":{"ref_index":4,"raw_excerpt":"Gong, E.J., Bang, C.S., Lee, J.J., Baik, G.H.: Knowledge-practice performance gap in clinical large language models: Systematic review of 39 benchmarks. Journal of medical Internet research 27, e84120 (2025). https://doi.org/10.2196/84120","verdict_class":"incontrovertible","doi_as_printed":"10.2196/84120"},"severity":"critical","ref_index":4,"audited_at":"2026-07-16T04:24:44.885804Z","event_type":"pith.integrity.v1","detected_doi":"10.2196/84120","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"broken_identifier","evidence_hash":"a1f7ac02c344eafd6a1c501b59c461e3b528f5209ec30cbea13dacc950a95338","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.0.0","detected_arxiv_id":null,"integrity_event_id":13427,"payload_sha256":"77d43ea35ce5e68f18bbd4a80af2dd9123021ed185bac2a4d13ee5e5b9ad2088","signature_b64":"yyofCbn+mOh9sJ/kytljJxoOiApD1vhCOzUxNwSiiEtIqKLAQ4lZE3RIxMvZkNVpHcBDKtH2dtWK7Z1aBED8CQ==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-16T04:29:08Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"coDpywRlQraiXnO0BZF+5ZGh6xIHnojHQnfz7A52zXBdVN61Q7uspZcAuHmTpXmI/s5ENQ9h7nfX+aKymg0GCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T17:12:31.890275Z"},"content_sha256":"25c1af23c9bfc98fea4e2000ad06c6072b4021723f90814137b5e7c6ac5e962e","schema_version":"1.0","event_id":"sha256:25c1af23c9bfc98fea4e2000ad06c6072b4021723f90814137b5e7c6ac5e962e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/IT2GD6TBP23FJ7VBXU7JLEFEVZ/bundle.json","state_url":"https://pith.science/pith/IT2GD6TBP23FJ7VBXU7JLEFEVZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/IT2GD6TBP23FJ7VBXU7JLEFEVZ/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-05T17:12:31Z","links":{"resolver":"https://pith.science/pith/IT2GD6TBP23FJ7VBXU7JLEFEVZ","bundle":"https://pith.science/pith/IT2GD6TBP23FJ7VBXU7JLEFEVZ/bundle.json","state":"https://pith.science/pith/IT2GD6TBP23FJ7VBXU7JLEFEVZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/IT2GD6TBP23FJ7VBXU7JLEFEVZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:IT2GD6TBP23FJ7VBXU7JLEFEVZ","merge_version":"pith-open-graph-merge-v1","event_count":3,"valid_event_count":3,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"d09fc8c21821920d28a77aea3098b1d168552b2ffc520bf6c6b35160d401b76c","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-10T12:04:31Z","title_canon_sha256":"648c9f0bf81196fc583ad4f319bf7de1ee1df0ff480596ec2fe802f6ae9cab93"},"schema_version":"1.0","source":{"id":"2607.09322","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.09322","created_at":"2026-07-13T01:19:02Z"},{"alias_kind":"arxiv_version","alias_value":"2607.09322v1","created_at":"2026-07-13T01:19:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.09322","created_at":"2026-07-13T01:19:02Z"},{"alias_kind":"pith_short_12","alias_value":"IT2GD6TBP23F","created_at":"2026-07-13T01:19:02Z"},{"alias_kind":"pith_short_16","alias_value":"IT2GD6TBP23FJ7VB","created_at":"2026-07-13T01:19:02Z"},{"alias_kind":"pith_short_8","alias_value":"IT2GD6TB","created_at":"2026-07-13T01:19:02Z"}],"graph_snapshots":[{"event_id":"sha256:4005db8e63aaaade59c167d5a3dc946668e60517da231e73d164a301f546f956","target":"graph","created_at":"2026-07-13T01:19:02Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2607.09322/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"In this work, we introduce LongMedBench, a real-world EHR-based benchmark for long-horizon clinical decision-making. Prior evaluations of LLM-based medical agents have largely emphasized short-context knowledge QA and tool use. However, real-world medical care is inherently longitudinal, and clinicians must aggregate evidence across repeated visits, tests, and evolving treatments. Therefore, long-horizon interaction is essential for realistic assessment. LongMedBench is constructed via a reproducible pipeline that integrates MIMIC-IV admission records and clinical notes into time-series event ","authors_text":"Hongxia Xu, Weiqi Zhai, Xiaocheng Zhang, Yanzhen Chen, Zhiting Fan, Zihan Xu, Zuozhu Liu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-10T12:04:31Z","title":"LongMedBench: Benchmarking Medical Agents for Long-Horizon Clinical Decision-Making"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.09322","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:b27a9b8149a8966517ae61f6c2a73c0bb7c3a674699b348d365330d9145931ac","target":"record","created_at":"2026-07-13T01:19:02Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"d09fc8c21821920d28a77aea3098b1d168552b2ffc520bf6c6b35160d401b76c","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-10T12:04:31Z","title_canon_sha256":"648c9f0bf81196fc583ad4f319bf7de1ee1df0ff480596ec2fe802f6ae9cab93"},"schema_version":"1.0","source":{"id":"2607.09322","kind":"arxiv","version":1}},"canonical_sha256":"44f461fa617eb654fea1bd3e9590a4ae41d437cc21cc9413af6e68fd773c1bc1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"44f461fa617eb654fea1bd3e9590a4ae41d437cc21cc9413af6e68fd773c1bc1","first_computed_at":"2026-07-13T01:19:02.338243Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-13T01:19:02.338243Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"lss3Ibn2J2N9snwE9Pp+ohaLrPNDUCtOvixXwvQGIUaM0qiDYICKfJi4arC1NP1suxPQKMmeQWyOPKSCvCHVBg==","signature_status":"signed_v1","signed_at":"2026-07-13T01:19:02.339414Z","signed_message":"canonical_sha256_bytes"},"source_id":"2607.09322","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b27a9b8149a8966517ae61f6c2a73c0bb7c3a674699b348d365330d9145931ac","sha256:4005db8e63aaaade59c167d5a3dc946668e60517da231e73d164a301f546f956","sha256:25c1af23c9bfc98fea4e2000ad06c6072b4021723f90814137b5e7c6ac5e962e"],"state_sha256":"1f073934821e2e9b5bd03fc689bd565231241d39660f8f0c5f6ce48d4a1a40e2"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZVK9em2Dpdpg5Sj/yxHJqqyZazwj0y3RC3vyawiSuX8UAt+TWDSHXCJldQArsTAb5NdvuQ9d6oTsfpCiey2bCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T17:12:31.892616Z","bundle_sha256":"318da97db1f4c4e065fa7127c91dc9efb88c0f29e9cf60aaa49fe2cc64508804"}}