{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:LBSTRZCLCFJRCYJ3PNDM3QWGQQ","short_pith_number":"pith:LBSTRZCL","canonical_record":{"source":{"id":"2503.04691","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-06T18:35:39Z","cross_cats_sorted":[],"title_canon_sha256":"5b685375c830393002b84d09fbf7e192f013cdc3c8fd439df12f71661a7fe945","abstract_canon_sha256":"48ee0e488769caa498c160d206cdfe41f10201f41411a76df054ff52d2ec290a"},"schema_version":"1.0"},"canonical_sha256":"586538e44b115311613b7b46cdc2c6843fd4d5bfe743e587a360256619aa7431","source":{"kind":"arxiv","id":"2503.04691","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.04691","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"arxiv_version","alias_value":"2503.04691v2","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.04691","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_12","alias_value":"LBSTRZCLCFJR","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_16","alias_value":"LBSTRZCLCFJRCYJ3","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_8","alias_value":"LBSTRZCL","created_at":"2026-07-05T10:28:05Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:LBSTRZCLCFJRCYJ3PNDM3QWGQQ","target":"record","payload":{"canonical_record":{"source":{"id":"2503.04691","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-06T18:35:39Z","cross_cats_sorted":[],"title_canon_sha256":"5b685375c830393002b84d09fbf7e192f013cdc3c8fd439df12f71661a7fe945","abstract_canon_sha256":"48ee0e488769caa498c160d206cdfe41f10201f41411a76df054ff52d2ec290a"},"schema_version":"1.0"},"canonical_sha256":"586538e44b115311613b7b46cdc2c6843fd4d5bfe743e587a360256619aa7431","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:28:05.833292Z","signature_b64":"1c1/RmCuol+fWNmvRzsHKsA0/BshaAoeoqNminVUEyjSLOwHsBvcaVdKesZMGlr3kup4L6iwanhP48Wyuzt2Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"586538e44b115311613b7b46cdc2c6843fd4d5bfe743e587a360256619aa7431","last_reissued_at":"2026-07-05T10:28:05.832630Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:28:05.832630Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2503.04691","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:28:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qPWSD33WHNFmpxdiyphYFQ1Xnnbj093tVkNieKbTPuSCreXxjnahzEgnmrbsCrjZ6fn+WRyBDUbMAckzo7yqAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-01T15:47:30.133278Z"},"content_sha256":"3649860bd123493e0d83f4ad0dba8087c2f2afb63f505996288248e5d029329d","schema_version":"1.0","event_id":"sha256:3649860bd123493e0d83f4ad0dba8087c2f2afb63f505996288248e5d029329d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:LBSTRZCLCFJRCYJ3PNDM3QWGQQ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Quantifying the Reasoning Abilities of LLMs on Real-world Clinical Cases","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chaoyi Wu, Chuanjin Peng, Hongfei Gu, Pengcheng Qiu, Shuyu Liu, Weidi Xie, Weike Zhao, Yanfeng Wang, Ya Zhang, Zhuoxia Chen","submitted_at":"2025-03-06T18:35:39Z","abstract_excerpt":"Recent advancements in reasoning-enhanced large language models (LLMs), such as DeepSeek-R1 and OpenAI-o3, have demonstrated significant progress. However, their application in professional medical contexts remains underexplored, particularly in evaluating the quality of their reasoning processes alongside final outputs. Here, we introduce MedR-Bench, a benchmarking dataset of 1,453 structured patient cases, annotated with reasoning references derived from clinical case reports. Spanning 13 body systems and 10 specialties, it includes both common and rare diseases. To comprehensively evaluate "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.04691","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/2503.04691/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":"2026-07-05T10:28:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QtzUsqakIJrIt0HwD0eEG96dICZ8Vw8xbRQcAjlMCWkITGzYrKX9HtjQLaIPxHgBFLwQKOhnjPszVGruJE1xBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-01T15:47:30.133757Z"},"content_sha256":"6296f1ee36a32b3b2c27b0417924ee57268669bef045a8409860ad47ce496cf4","schema_version":"1.0","event_id":"sha256:6296f1ee36a32b3b2c27b0417924ee57268669bef045a8409860ad47ce496cf4"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LBSTRZCLCFJRCYJ3PNDM3QWGQQ/bundle.json","state_url":"https://pith.science/pith/LBSTRZCLCFJRCYJ3PNDM3QWGQQ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LBSTRZCLCFJRCYJ3PNDM3QWGQQ/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":"2026-08-01T15:47:30Z","links":{"resolver":"https://pith.science/pith/LBSTRZCLCFJRCYJ3PNDM3QWGQQ","bundle":"https://pith.science/pith/LBSTRZCLCFJRCYJ3PNDM3QWGQQ/bundle.json","state":"https://pith.science/pith/LBSTRZCLCFJRCYJ3PNDM3QWGQQ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LBSTRZCLCFJRCYJ3PNDM3QWGQQ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:LBSTRZCLCFJRCYJ3PNDM3QWGQQ","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"48ee0e488769caa498c160d206cdfe41f10201f41411a76df054ff52d2ec290a","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-06T18:35:39Z","title_canon_sha256":"5b685375c830393002b84d09fbf7e192f013cdc3c8fd439df12f71661a7fe945"},"schema_version":"1.0","source":{"id":"2503.04691","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.04691","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"arxiv_version","alias_value":"2503.04691v2","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.04691","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_12","alias_value":"LBSTRZCLCFJR","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_16","alias_value":"LBSTRZCLCFJRCYJ3","created_at":"2026-07-05T10:28:05Z"},{"alias_kind":"pith_short_8","alias_value":"LBSTRZCL","created_at":"2026-07-05T10:28:05Z"}],"graph_snapshots":[{"event_id":"sha256:6296f1ee36a32b3b2c27b0417924ee57268669bef045a8409860ad47ce496cf4","target":"graph","created_at":"2026-07-05T10:28:05Z","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/2503.04691/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent advancements in reasoning-enhanced large language models (LLMs), such as DeepSeek-R1 and OpenAI-o3, have demonstrated significant progress. However, their application in professional medical contexts remains underexplored, particularly in evaluating the quality of their reasoning processes alongside final outputs. Here, we introduce MedR-Bench, a benchmarking dataset of 1,453 structured patient cases, annotated with reasoning references derived from clinical case reports. Spanning 13 body systems and 10 specialties, it includes both common and rare diseases. To comprehensively evaluate ","authors_text":"Chaoyi Wu, Chuanjin Peng, Hongfei Gu, Pengcheng Qiu, Shuyu Liu, Weidi Xie, Weike Zhao, Yanfeng Wang, Ya Zhang, Zhuoxia Chen","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-06T18:35:39Z","title":"Quantifying the Reasoning Abilities of LLMs on Real-world Clinical Cases"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.04691","kind":"arxiv","version":2},"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":"sha256:3649860bd123493e0d83f4ad0dba8087c2f2afb63f505996288248e5d029329d","target":"record","created_at":"2026-07-05T10:28:05Z","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":"48ee0e488769caa498c160d206cdfe41f10201f41411a76df054ff52d2ec290a","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-06T18:35:39Z","title_canon_sha256":"5b685375c830393002b84d09fbf7e192f013cdc3c8fd439df12f71661a7fe945"},"schema_version":"1.0","source":{"id":"2503.04691","kind":"arxiv","version":2}},"canonical_sha256":"586538e44b115311613b7b46cdc2c6843fd4d5bfe743e587a360256619aa7431","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"586538e44b115311613b7b46cdc2c6843fd4d5bfe743e587a360256619aa7431","first_computed_at":"2026-07-05T10:28:05.832630Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:28:05.832630Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"1c1/RmCuol+fWNmvRzsHKsA0/BshaAoeoqNminVUEyjSLOwHsBvcaVdKesZMGlr3kup4L6iwanhP48Wyuzt2Aw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:28:05.833292Z","signed_message":"canonical_sha256_bytes"},"source_id":"2503.04691","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3649860bd123493e0d83f4ad0dba8087c2f2afb63f505996288248e5d029329d","sha256:6296f1ee36a32b3b2c27b0417924ee57268669bef045a8409860ad47ce496cf4"],"state_sha256":"fb2a2ba03b464dc1802ba13e3b307ce2281f84d59866f947107a3a7c9ad20cb8"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rufmmTMqEe425KG7IPgYbA/6iy3tvEftxZDOBFiZwPiNq5AniKYN9qKWEYLvobU9T6X8xfitRywGJAhvYurRDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-01T15:47:30.137247Z","bundle_sha256":"04d2a33db1c37d2d1e2df6225b8850fd0b439f783767d0077e3007641be9931c"}}