{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:IZ3TAFUZJYUHFJLKL64CJIHBMD","short_pith_number":"pith:IZ3TAFUZ","canonical_record":{"source":{"id":"2605.17034","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T15:08:31Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"5d685e6cab6fdeae10a62320a57a8f0618604a2ba988756010cb426660087607","abstract_canon_sha256":"ca655135676dec20c6b41f22d68516ed5632475ac4cbe76b8ab2475555d7f230"},"schema_version":"1.0"},"canonical_sha256":"46773016994e2872a56a5fb824a0e160e77aa961d5f45f32c38c4c84c8e64ef2","source":{"kind":"arxiv","id":"2605.17034","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.17034","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"arxiv_version","alias_value":"2605.17034v1","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17034","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"pith_short_12","alias_value":"IZ3TAFUZJYUH","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"pith_short_16","alias_value":"IZ3TAFUZJYUHFJLK","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"pith_short_8","alias_value":"IZ3TAFUZ","created_at":"2026-05-20T00:03:36Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:IZ3TAFUZJYUHFJLKL64CJIHBMD","target":"record","payload":{"canonical_record":{"source":{"id":"2605.17034","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T15:08:31Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"5d685e6cab6fdeae10a62320a57a8f0618604a2ba988756010cb426660087607","abstract_canon_sha256":"ca655135676dec20c6b41f22d68516ed5632475ac4cbe76b8ab2475555d7f230"},"schema_version":"1.0"},"canonical_sha256":"46773016994e2872a56a5fb824a0e160e77aa961d5f45f32c38c4c84c8e64ef2","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:03:36.956197Z","signature_b64":"ZygkynsaPkt1s/PgB8ijxI8YNWcWzY2XzhePsCm4Nnj1VzCpZUZUSL0/P7YUXuespN6iQEY3jqBZJRiJzIf7DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"46773016994e2872a56a5fb824a0e160e77aa961d5f45f32c38c4c84c8e64ef2","last_reissued_at":"2026-05-20T00:03:36.955421Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:03:36.955421Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.17034","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":"2026-05-20T00:03:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4Og16hg7bCuT6XIynwOZcivy33tgJmblHLQDlX5hJv9nXQkr6dqd0iE+UODohDk+HH51uKY08rbNACqMGNbbBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-09T09:03:47.438935Z"},"content_sha256":"6383dd70af8de29575934dd69f8587611cd2c4c4f69fd41c535e298b89b8bb7f","schema_version":"1.0","event_id":"sha256:6383dd70af8de29575934dd69f8587611cd2c4c4f69fd41c535e298b89b8bb7f"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:IZ3TAFUZJYUHFJLKL64CJIHBMD","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Privacy Policy Enforcement Guardrails for Data-Sensitive Retrieval-Augmented Generation","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"Dual one-class density estimators detect contextual privacy leaks in RAG with over 0.93 AUROC on borderline cases.","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.LG","authors_text":"Alexander Nemecek, Debargha Ganguly, Erman Ayday, Osama Zafar, Vikash Singh, Vipin Chaudhary, Wenbiao Li, Yiqian Zhang","submitted_at":"2026-05-16T15:08:31Z","abstract_excerpt":"Standard PII filters often miss contextual data leakage in RAG systems, such as non-regulated attribute clusters that collectively identify individuals. We introduce a Privacy Policy Enforcement (PPE) framework using dual one-class density estimators with fused text embeddings and a calibrated abstain region for out-of-distribution inputs. Using an axis-stratified, multi-LLM synthetic data pipeline across medicine, finance, and law, we found that traditional Gaussian Mixture baselines fail on borderline-safe stress tests by focusing on linguistic register rather than content.\n  Our proposed T3"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our proposed T3+OCSVM detector, trained on safe and borderline-safe data, achieves a borderline AUROC of 0.93+ while reducing false positives by 44-55 percentage points and maintaining millisecond latency.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The axis-stratified multi-LLM synthetic data pipeline produces borderline-safe examples whose privacy leakage properties match those encountered in real RAG deployments across medicine, finance, and law.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Presents T3+OCSVM detector for privacy policy enforcement in RAG achieving 0.93+ borderline AUROC, 44-55 point false positive reduction, and millisecond latency via synthetic data stress tests.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Dual one-class density estimators detect contextual privacy leaks in RAG with over 0.93 AUROC on borderline cases.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"80b5ecb2bba23f649de1fbc2f126aba74c0cd73c36c6f5e949714cf808505700"},"source":{"id":"2605.17034","kind":"arxiv","version":1},"verdict":{"id":"4abbb724-12e1-44ee-b6a6-27afadba7f69","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T20:22:58.470375Z","strongest_claim":"Our proposed T3+OCSVM detector, trained on safe and borderline-safe data, achieves a borderline AUROC of 0.93+ while reducing false positives by 44-55 percentage points and maintaining millisecond latency.","one_line_summary":"Presents T3+OCSVM detector for privacy policy enforcement in RAG achieving 0.93+ borderline AUROC, 44-55 point false positive reduction, and millisecond latency via synthetic data stress tests.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The axis-stratified multi-LLM synthetic data pipeline produces borderline-safe examples whose privacy leakage properties match those encountered in real RAG deployments across medicine, finance, and law.","pith_extraction_headline":"Dual one-class density estimators detect contextual privacy leaks in RAG with over 0.93 AUROC on borderline cases."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.17034/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T20:31:18.951259Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T20:30:42.475170Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"citation_quote_validity","ran_at":"2026-05-19T19:49:42.694997Z","status":"skipped","version":"0.1.0","findings_count":0},{"name":"cited_work_retraction","ran_at":"2026-05-19T18:51:56.821928Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.169455Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:23.186185Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"0193efc6fca9b1ef231c5c252d4b93ec7fdb0dcbd24c53a607890701e2538316"},"references":{"count":53,"sample":[{"doi":"","year":2026,"title":"D. Gangulyet al., “Trust the typical,”arXiv preprint arXiv:2602.04581, 2026","work_id":"f9953d8e-eb2a-4e8e-9610-1df1dda10d04","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2002,"title":"k-anonymity: A model for protecting privacy,","work_id":"9c1101af-3a6b-4976-9854-b1a37b1eef8f","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"Extracting training data from large language models,","work_id":"bd432c94-7950-4690-8ecd-c37b4d439fe4","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Quantifying memorization across neural language models,","work_id":"124d08c5-1b98-4d7e-8b8b-bb3f32fb1772","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Scalable Extraction of Training Data from (Production) Language Models","work_id":"7ee4de98-0bdd-47ab-abe6-1865cb65b1ae","ref_index":5,"cited_arxiv_id":"2311.17035","is_internal_anchor":true}],"resolved_work":53,"snapshot_sha256":"448aeaa05e548695a299ebc82af77e596b26f8a9748db8d687c4fac6e455be8f","internal_anchors":6},"formal_canon":{"evidence_count":2,"snapshot_sha256":"88ea6a6fc4691978c64ef60a41009ed9326d3ed1810952ba40c2834c5f3d807b"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"4abbb724-12e1-44ee-b6a6-27afadba7f69"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:03:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"y578iQyPdX2tPTtoy6Tp0AFBo/ZN0QNYQ1saf3XqWnr7FNajpEG3yybJYVHoEqlMRwX2Vuev4Pqs1hLGR0/QDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-09T09:03:47.440330Z"},"content_sha256":"e874a3b30bf3aa871d6f66a5354f611d7667911b06fa77bbf4222bceea93b728","schema_version":"1.0","event_id":"sha256:e874a3b30bf3aa871d6f66a5354f611d7667911b06fa77bbf4222bceea93b728"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/IZ3TAFUZJYUHFJLKL64CJIHBMD/bundle.json","state_url":"https://pith.science/pith/IZ3TAFUZJYUHFJLKL64CJIHBMD/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/IZ3TAFUZJYUHFJLKL64CJIHBMD/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":"2026-06-09T09:03:47Z","links":{"resolver":"https://pith.science/pith/IZ3TAFUZJYUHFJLKL64CJIHBMD","bundle":"https://pith.science/pith/IZ3TAFUZJYUHFJLKL64CJIHBMD/bundle.json","state":"https://pith.science/pith/IZ3TAFUZJYUHFJLKL64CJIHBMD/state.json","well_known_bundle":"https://pith.science/.well-known/pith/IZ3TAFUZJYUHFJLKL64CJIHBMD/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:IZ3TAFUZJYUHFJLKL64CJIHBMD","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":"ca655135676dec20c6b41f22d68516ed5632475ac4cbe76b8ab2475555d7f230","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T15:08:31Z","title_canon_sha256":"5d685e6cab6fdeae10a62320a57a8f0618604a2ba988756010cb426660087607"},"schema_version":"1.0","source":{"id":"2605.17034","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.17034","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"arxiv_version","alias_value":"2605.17034v1","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17034","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"pith_short_12","alias_value":"IZ3TAFUZJYUH","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"pith_short_16","alias_value":"IZ3TAFUZJYUHFJLK","created_at":"2026-05-20T00:03:36Z"},{"alias_kind":"pith_short_8","alias_value":"IZ3TAFUZ","created_at":"2026-05-20T00:03:36Z"}],"graph_snapshots":[{"event_id":"sha256:e874a3b30bf3aa871d6f66a5354f611d7667911b06fa77bbf4222bceea93b728","target":"graph","created_at":"2026-05-20T00:03:36Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Our proposed T3+OCSVM detector, trained on safe and borderline-safe data, achieves a borderline AUROC of 0.93+ while reducing false positives by 44-55 percentage points and maintaining millisecond latency."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The axis-stratified multi-LLM synthetic data pipeline produces borderline-safe examples whose privacy leakage properties match those encountered in real RAG deployments across medicine, finance, and law."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Presents T3+OCSVM detector for privacy policy enforcement in RAG achieving 0.93+ borderline AUROC, 44-55 point false positive reduction, and millisecond latency via synthetic data stress tests."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Dual one-class density estimators detect contextual privacy leaks in RAG with over 0.93 AUROC on borderline cases."}],"snapshot_sha256":"80b5ecb2bba23f649de1fbc2f126aba74c0cd73c36c6f5e949714cf808505700"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"88ea6a6fc4691978c64ef60a41009ed9326d3ed1810952ba40c2834c5f3d807b"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T20:31:18.951259Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T20:30:42.475170Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"citation_quote_validity","ran_at":"2026-05-19T19:49:42.694997Z","status":"skipped","version":"0.1.0"},{"findings_count":0,"name":"cited_work_retraction","ran_at":"2026-05-19T18:51:56.821928Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.169455Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:23.186185Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.17034/integrity.json","findings":[],"snapshot_sha256":"0193efc6fca9b1ef231c5c252d4b93ec7fdb0dcbd24c53a607890701e2538316","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Standard PII filters often miss contextual data leakage in RAG systems, such as non-regulated attribute clusters that collectively identify individuals. We introduce a Privacy Policy Enforcement (PPE) framework using dual one-class density estimators with fused text embeddings and a calibrated abstain region for out-of-distribution inputs. Using an axis-stratified, multi-LLM synthetic data pipeline across medicine, finance, and law, we found that traditional Gaussian Mixture baselines fail on borderline-safe stress tests by focusing on linguistic register rather than content.\n  Our proposed T3","authors_text":"Alexander Nemecek, Debargha Ganguly, Erman Ayday, Osama Zafar, Vikash Singh, Vipin Chaudhary, Wenbiao Li, Yiqian Zhang","cross_cats":["cs.AI","cs.CR"],"headline":"Dual one-class density estimators detect contextual privacy leaks in RAG with over 0.93 AUROC on borderline cases.","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T15:08:31Z","title":"Privacy Policy Enforcement Guardrails for Data-Sensitive Retrieval-Augmented Generation"},"references":{"count":53,"internal_anchors":6,"resolved_work":53,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"D. Gangulyet al., “Trust the typical,”arXiv preprint arXiv:2602.04581, 2026","work_id":"f9953d8e-eb2a-4e8e-9610-1df1dda10d04","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"k-anonymity: A model for protecting privacy,","work_id":"9c1101af-3a6b-4976-9854-b1a37b1eef8f","year":2002},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Extracting training data from large language models,","work_id":"bd432c94-7950-4690-8ecd-c37b4d439fe4","year":2021},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Quantifying memorization across neural language models,","work_id":"124d08c5-1b98-4d7e-8b8b-bb3f32fb1772","year":2023},{"cited_arxiv_id":"2311.17035","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Scalable Extraction of Training Data from (Production) Language Models","work_id":"7ee4de98-0bdd-47ab-abe6-1865cb65b1ae","year":2023}],"snapshot_sha256":"448aeaa05e548695a299ebc82af77e596b26f8a9748db8d687c4fac6e455be8f"},"source":{"id":"2605.17034","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T20:22:58.470375Z","id":"4abbb724-12e1-44ee-b6a6-27afadba7f69","model_set":{"reader":"grok-4.3"},"one_line_summary":"Presents T3+OCSVM detector for privacy policy enforcement in RAG achieving 0.93+ borderline AUROC, 44-55 point false positive reduction, and millisecond latency via synthetic data stress tests.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Dual one-class density estimators detect contextual privacy leaks in RAG with over 0.93 AUROC on borderline cases.","strongest_claim":"Our proposed T3+OCSVM detector, trained on safe and borderline-safe data, achieves a borderline AUROC of 0.93+ while reducing false positives by 44-55 percentage points and maintaining millisecond latency.","weakest_assumption":"The axis-stratified multi-LLM synthetic data pipeline produces borderline-safe examples whose privacy leakage properties match those encountered in real RAG deployments across medicine, finance, and law."}},"verdict_id":"4abbb724-12e1-44ee-b6a6-27afadba7f69"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:6383dd70af8de29575934dd69f8587611cd2c4c4f69fd41c535e298b89b8bb7f","target":"record","created_at":"2026-05-20T00:03:36Z","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":"ca655135676dec20c6b41f22d68516ed5632475ac4cbe76b8ab2475555d7f230","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T15:08:31Z","title_canon_sha256":"5d685e6cab6fdeae10a62320a57a8f0618604a2ba988756010cb426660087607"},"schema_version":"1.0","source":{"id":"2605.17034","kind":"arxiv","version":1}},"canonical_sha256":"46773016994e2872a56a5fb824a0e160e77aa961d5f45f32c38c4c84c8e64ef2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"46773016994e2872a56a5fb824a0e160e77aa961d5f45f32c38c4c84c8e64ef2","first_computed_at":"2026-05-20T00:03:36.955421Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:03:36.955421Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ZygkynsaPkt1s/PgB8ijxI8YNWcWzY2XzhePsCm4Nnj1VzCpZUZUSL0/P7YUXuespN6iQEY3jqBZJRiJzIf7DA==","signature_status":"signed_v1","signed_at":"2026-05-20T00:03:36.956197Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.17034","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6383dd70af8de29575934dd69f8587611cd2c4c4f69fd41c535e298b89b8bb7f","sha256:e874a3b30bf3aa871d6f66a5354f611d7667911b06fa77bbf4222bceea93b728"],"state_sha256":"47a147fb106c551a965055503f386148561b7cdcddc282072e4021c9ff379d52"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Cvy1oJUSQ5n4g4PsF/gdClaqGBW8guNm3ofhCoAkA3dxcc3qpiVjzXnYeDZJmxgQe8UGALkegg0INcLYLmE8Dg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-09T09:03:47.446636Z","bundle_sha256":"6fbbdff1185a099445ea33d5160a3cfaee9fc3303440f0edf5b662461b9010da"}}