{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:JO66NCQVVZTZHEZSXKAR753EWP","short_pith_number":"pith:JO66NCQV","schema_version":"1.0","canonical_sha256":"4bbde68a15ae67939332ba811ff764b3dfa6b767eba8fefb21dc777e1905a97e","source":{"kind":"arxiv","id":"2607.01470","version":1},"attestation_state":"computed","paper":{"title":"World Feedback for Clinical Agents: Diagnosing RL in FHIR Environments","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Abhishek Mukherji, Ananya Mantravadi, Harshit Rajgarhia, Prasanna Desikan","submitted_at":"2026-07-01T21:02:54Z","abstract_excerpt":"Clinical protocol-execution tasks -- checking a lab value, applying a threshold, placing a correctly structured FHIR order -- are natural candidates for RL from world feedback: once clinical SMEs encode decision logic into a verifier, that verifier grades unlimited rollouts without per-episode annotation. But applying RL requires a sound feedback channel and sufficient base capability. We audit MedAgentBench v1/v2, find a 41.7\\% silent-finish ceiling that makes inaction the RL dominant strategy, and construct \\textbf{MedAgentBench-v3 (MAB-v3)} (508 tasks, 8.9\\% ceiling). Training Qwen3-8B expo"},"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":"2607.01470","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-01T21:02:54Z","cross_cats_sorted":[],"title_canon_sha256":"ff50a7bdfba3b43b95d4580f29f83dd0c35f5868e6261e53bfb6e5930ed3e8d0","abstract_canon_sha256":"e42601ce5856c5a0a7007ef1da86210c236c86e5bee4b80c70b8482cad250fc6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-03T00:17:01.070496Z","signature_b64":"xnQlFmv3HP1uz6RgCIIWZ0PK4wCn7SoKI/f7HSuK3jmJfRNtpgulZ+jsd8J7E3/SAy6PRiu6m+YtKAMSJflqBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4bbde68a15ae67939332ba811ff764b3dfa6b767eba8fefb21dc777e1905a97e","last_reissued_at":"2026-07-03T00:17:01.070088Z","signature_status":"signed_v1","first_computed_at":"2026-07-03T00:17:01.070088Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"World Feedback for Clinical Agents: Diagnosing RL in FHIR Environments","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Abhishek Mukherji, Ananya Mantravadi, Harshit Rajgarhia, Prasanna Desikan","submitted_at":"2026-07-01T21:02:54Z","abstract_excerpt":"Clinical protocol-execution tasks -- checking a lab value, applying a threshold, placing a correctly structured FHIR order -- are natural candidates for RL from world feedback: once clinical SMEs encode decision logic into a verifier, that verifier grades unlimited rollouts without per-episode annotation. But applying RL requires a sound feedback channel and sufficient base capability. We audit MedAgentBench v1/v2, find a 41.7\\% silent-finish ceiling that makes inaction the RL dominant strategy, and construct \\textbf{MedAgentBench-v3 (MAB-v3)} (508 tasks, 8.9\\% ceiling). Training Qwen3-8B expo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.01470","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/2607.01470/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":"2607.01470","created_at":"2026-07-03T00:17:01.070145+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.01470v1","created_at":"2026-07-03T00:17:01.070145+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.01470","created_at":"2026-07-03T00:17:01.070145+00:00"},{"alias_kind":"pith_short_12","alias_value":"JO66NCQVVZTZ","created_at":"2026-07-03T00:17:01.070145+00:00"},{"alias_kind":"pith_short_16","alias_value":"JO66NCQVVZTZHEZS","created_at":"2026-07-03T00:17:01.070145+00:00"},{"alias_kind":"pith_short_8","alias_value":"JO66NCQV","created_at":"2026-07-03T00:17:01.070145+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JO66NCQVVZTZHEZSXKAR753EWP","json":"https://pith.science/pith/JO66NCQVVZTZHEZSXKAR753EWP.json","graph_json":"https://pith.science/api/pith-number/JO66NCQVVZTZHEZSXKAR753EWP/graph.json","events_json":"https://pith.science/api/pith-number/JO66NCQVVZTZHEZSXKAR753EWP/events.json","paper":"https://pith.science/paper/JO66NCQV"},"agent_actions":{"view_html":"https://pith.science/pith/JO66NCQVVZTZHEZSXKAR753EWP","download_json":"https://pith.science/pith/JO66NCQVVZTZHEZSXKAR753EWP.json","view_paper":"https://pith.science/paper/JO66NCQV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.01470&json=true","fetch_graph":"https://pith.science/api/pith-number/JO66NCQVVZTZHEZSXKAR753EWP/graph.json","fetch_events":"https://pith.science/api/pith-number/JO66NCQVVZTZHEZSXKAR753EWP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JO66NCQVVZTZHEZSXKAR753EWP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JO66NCQVVZTZHEZSXKAR753EWP/action/storage_attestation","attest_author":"https://pith.science/pith/JO66NCQVVZTZHEZSXKAR753EWP/action/author_attestation","sign_citation":"https://pith.science/pith/JO66NCQVVZTZHEZSXKAR753EWP/action/citation_signature","submit_replication":"https://pith.science/pith/JO66NCQVVZTZHEZSXKAR753EWP/action/replication_record"}},"created_at":"2026-07-03T00:17:01.070145+00:00","updated_at":"2026-07-03T00:17:01.070145+00:00"}