{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:QBQNMUWYUFGFTG7XS2YJ4SF3E2","short_pith_number":"pith:QBQNMUWY","canonical_record":{"source":{"id":"2605.17267","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2026-05-17T05:21:23Z","cross_cats_sorted":[],"title_canon_sha256":"0da7fb68f4cf3afead675656b05cbbde0f33099287aaa483aea79ec6c3adb077","abstract_canon_sha256":"beaed6b8b70a1ce6254095d23e067deb0778ddae9ebc246ae260e53d6be7db56"},"schema_version":"1.0"},"canonical_sha256":"8060d652d8a14c599bf796b09e48bb26b077e3845c508e042f8c6d7596a45385","source":{"kind":"arxiv","id":"2605.17267","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.17267","created_at":"2026-05-20T00:03:48Z"},{"alias_kind":"arxiv_version","alias_value":"2605.17267v1","created_at":"2026-05-20T00:03:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17267","created_at":"2026-05-20T00:03:48Z"},{"alias_kind":"pith_short_12","alias_value":"QBQNMUWYUFGF","created_at":"2026-05-20T00:03:48Z"},{"alias_kind":"pith_short_16","alias_value":"QBQNMUWYUFGFTG7X","created_at":"2026-05-20T00:03:48Z"},{"alias_kind":"pith_short_8","alias_value":"QBQNMUWY","created_at":"2026-05-20T00:03:48Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:QBQNMUWYUFGFTG7XS2YJ4SF3E2","target":"record","payload":{"canonical_record":{"source":{"id":"2605.17267","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2026-05-17T05:21:23Z","cross_cats_sorted":[],"title_canon_sha256":"0da7fb68f4cf3afead675656b05cbbde0f33099287aaa483aea79ec6c3adb077","abstract_canon_sha256":"beaed6b8b70a1ce6254095d23e067deb0778ddae9ebc246ae260e53d6be7db56"},"schema_version":"1.0"},"canonical_sha256":"8060d652d8a14c599bf796b09e48bb26b077e3845c508e042f8c6d7596a45385","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:03:48.865153Z","signature_b64":"fFpw69ywGIhPnPVUOsKeVTruMQ9FtJNC8pPQoCxzE6pEU79NKrg1L3BNDk5N91wswfR8SU7ovFE1SHwLIdkaBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8060d652d8a14c599bf796b09e48bb26b077e3845c508e042f8c6d7596a45385","last_reissued_at":"2026-05-20T00:03:48.864260Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:03:48.864260Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.17267","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":"2026-05-20T00:03:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"a5UczTMT60QmMrbbEUeKu5gRXkhA+ZAs0p8RkY2IIn2MfL2CWwZeb3R+8RM0FS04ztghYaFs3DqamGONkytUBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-29T18:00:12.149928Z"},"content_sha256":"a7179449eaff191ad9afebf21a30847a4cfecfdafb27b6facc56a9cf50ad8bed","schema_version":"1.0","event_id":"sha256:a7179449eaff191ad9afebf21a30847a4cfecfdafb27b6facc56a9cf50ad8bed"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:QBQNMUWYUFGFTG7XS2YJ4SF3E2","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"RAGR: Review-Augmented Generative Recommendation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Interleaving review semantic IDs into item sequences lets generative models use explanatory feedback to improve next-item predictions.","cross_cats":[],"primary_cat":"cs.IR","authors_text":"Junyi Li, Sheng Zhang, Wenlin Zhang, Xiangyu Zhao, Xianneng Li, Xiaowei Qian, Yejing Wang, Yichao Wang, Yingyi Zhang, Yong Liu, Yue Feng","submitted_at":"2026-05-17T05:21:23Z","abstract_excerpt":"Sequential recommendation (SR) is traditionally formulated as next-item prediction over a chronological sequence of interacted items. Although recent generative recommendation (GR) methods introduce new machinery, such as semantic IDs, autoregressive decoding, and unified token spaces, they largely inherit the same item-only modeling assumption. We argue that this design constitutes a structural bottleneck, because user decision-making is not purely behavioral: while item interactions reveal what users choose, review feedback often explain why they choose it by exposing latent evaluative facto"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments on three real-world datasets show that RAGR yields consistent and significant gains over strong GR backbones across all metrics.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That interleaving review semantic IDs into the generative user sequence allows review signals to participate directly in autoregressive next-token generation while the Item-Centric Task Generation Alignment based on DPO successfully preserves the pure recommendation objective without introducing new biases or degrading item prediction quality.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"RAGR builds mixed item-review sequences for generative recommendation and uses DPO alignment to favor item tokens, reporting gains over prior GR baselines on three datasets.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Interleaving review semantic IDs into item sequences lets generative models use explanatory feedback to improve next-item predictions.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"46a0fcb7aa108d32ac153c32ce7544292e4558f24fbd50adebbfb6a2eb5dfcfa"},"source":{"id":"2605.17267","kind":"arxiv","version":1},"verdict":{"id":"5939590b-8fab-4eba-b0eb-975c19f96497","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T23:16:31.126607Z","strongest_claim":"Experiments on three real-world datasets show that RAGR yields consistent and significant gains over strong GR backbones across all metrics.","one_line_summary":"RAGR builds mixed item-review sequences for generative recommendation and uses DPO alignment to favor item tokens, reporting gains over prior GR baselines on three datasets.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That interleaving review semantic IDs into the generative user sequence allows review signals to participate directly in autoregressive next-token generation while the Item-Centric Task Generation Alignment based on DPO successfully preserves the pure recommendation objective without introducing new biases or degrading item prediction quality.","pith_extraction_headline":"Interleaving review semantic IDs into item sequences lets generative models use explanatory feedback to improve next-item predictions."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.17267/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T23:31:20.232082Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T23:31:00.864371Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T22:01:57.841057Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T21:33:23.780242Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"d9554c1b90a31784d30f0433e83abcf729fb000602a2775c9ae952305c2ab11b"},"references":{"count":51,"sample":[{"doi":"","year":2024,"title":"Recommender systems in the era of large language models (llms),","work_id":"4901ccf9-ca41-45fc-83fd-4175873ceca9","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2016,"title":"Ups and downs: Modeling the visual evolution of fashion trends with one-class collaborative filtering,","work_id":"31e853b0-438f-49bc-97e8-c80ff91a807f","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"Billion- scale commodity embedding for e-commerce recommendation in alibaba,","work_id":"4f1d6651-8a09-4bdd-8236-9d62b05ff329","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"W.-C. Kang and J. McAuley, “Self-attentive sequential recommendation,” in2018 IEEE international conference on data mining (ICDM). IEEE, 2018, pp. 197–206. 1, 5, 6, 11","work_id":"726c788a-79e8-4201-834b-be022e969ccc","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Contrastive learning for sequential recommendation,","work_id":"5ab05224-583d-4f76-bc70-9fd87386a327","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":51,"snapshot_sha256":"b1c6097f1cf2b944a00e3f57152c8f5708f2dc2489b057253ddfb6173415f623","internal_anchors":1},"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":"5939590b-8fab-4eba-b0eb-975c19f96497"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:03:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"gs2B4SzNIdmfrVrGCxvPXlM0grhYnH1oMmyJzgqGkheUHEpjSmT8h0v3RLS/VwPNKiAOM65xrIl+TMcjnPdLAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-29T18:00:12.151391Z"},"content_sha256":"ad95ec8a344dda252cba6c27c632f2c20b317f2527df9e72e3e9da61298e973e","schema_version":"1.0","event_id":"sha256:ad95ec8a344dda252cba6c27c632f2c20b317f2527df9e72e3e9da61298e973e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/QBQNMUWYUFGFTG7XS2YJ4SF3E2/bundle.json","state_url":"https://pith.science/pith/QBQNMUWYUFGFTG7XS2YJ4SF3E2/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/QBQNMUWYUFGFTG7XS2YJ4SF3E2/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":"2026-05-29T18:00:12Z","links":{"resolver":"https://pith.science/pith/QBQNMUWYUFGFTG7XS2YJ4SF3E2","bundle":"https://pith.science/pith/QBQNMUWYUFGFTG7XS2YJ4SF3E2/bundle.json","state":"https://pith.science/pith/QBQNMUWYUFGFTG7XS2YJ4SF3E2/state.json","well_known_bundle":"https://pith.science/.well-known/pith/QBQNMUWYUFGFTG7XS2YJ4SF3E2/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:QBQNMUWYUFGFTG7XS2YJ4SF3E2","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":"beaed6b8b70a1ce6254095d23e067deb0778ddae9ebc246ae260e53d6be7db56","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2026-05-17T05:21:23Z","title_canon_sha256":"0da7fb68f4cf3afead675656b05cbbde0f33099287aaa483aea79ec6c3adb077"},"schema_version":"1.0","source":{"id":"2605.17267","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.17267","created_at":"2026-05-20T00:03:48Z"},{"alias_kind":"arxiv_version","alias_value":"2605.17267v1","created_at":"2026-05-20T00:03:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17267","created_at":"2026-05-20T00:03:48Z"},{"alias_kind":"pith_short_12","alias_value":"QBQNMUWYUFGF","created_at":"2026-05-20T00:03:48Z"},{"alias_kind":"pith_short_16","alias_value":"QBQNMUWYUFGFTG7X","created_at":"2026-05-20T00:03:48Z"},{"alias_kind":"pith_short_8","alias_value":"QBQNMUWY","created_at":"2026-05-20T00:03:48Z"}],"graph_snapshots":[{"event_id":"sha256:ad95ec8a344dda252cba6c27c632f2c20b317f2527df9e72e3e9da61298e973e","target":"graph","created_at":"2026-05-20T00:03:48Z","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":"Experiments on three real-world datasets show that RAGR yields consistent and significant gains over strong GR backbones across all metrics."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That interleaving review semantic IDs into the generative user sequence allows review signals to participate directly in autoregressive next-token generation while the Item-Centric Task Generation Alignment based on DPO successfully preserves the pure recommendation objective without introducing new biases or degrading item prediction quality."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"RAGR builds mixed item-review sequences for generative recommendation and uses DPO alignment to favor item tokens, reporting gains over prior GR baselines on three datasets."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Interleaving review semantic IDs into item sequences lets generative models use explanatory feedback to improve next-item predictions."}],"snapshot_sha256":"46a0fcb7aa108d32ac153c32ce7544292e4558f24fbd50adebbfb6a2eb5dfcfa"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T23:31:20.232082Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T23:31:00.864371Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T22:01:57.841057Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T21:33:23.780242Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.17267/integrity.json","findings":[],"snapshot_sha256":"d9554c1b90a31784d30f0433e83abcf729fb000602a2775c9ae952305c2ab11b","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Sequential recommendation (SR) is traditionally formulated as next-item prediction over a chronological sequence of interacted items. Although recent generative recommendation (GR) methods introduce new machinery, such as semantic IDs, autoregressive decoding, and unified token spaces, they largely inherit the same item-only modeling assumption. We argue that this design constitutes a structural bottleneck, because user decision-making is not purely behavioral: while item interactions reveal what users choose, review feedback often explain why they choose it by exposing latent evaluative facto","authors_text":"Junyi Li, Sheng Zhang, Wenlin Zhang, Xiangyu Zhao, Xianneng Li, Xiaowei Qian, Yejing Wang, Yichao Wang, Yingyi Zhang, Yong Liu, Yue Feng","cross_cats":[],"headline":"Interleaving review semantic IDs into item sequences lets generative models use explanatory feedback to improve next-item predictions.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2026-05-17T05:21:23Z","title":"RAGR: Review-Augmented Generative Recommendation"},"references":{"count":51,"internal_anchors":1,"resolved_work":51,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Recommender systems in the era of large language models (llms),","work_id":"4901ccf9-ca41-45fc-83fd-4175873ceca9","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Ups and downs: Modeling the visual evolution of fashion trends with one-class collaborative filtering,","work_id":"31e853b0-438f-49bc-97e8-c80ff91a807f","year":2016},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Billion- scale commodity embedding for e-commerce recommendation in alibaba,","work_id":"4f1d6651-8a09-4bdd-8236-9d62b05ff329","year":2018},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"W.-C. Kang and J. McAuley, “Self-attentive sequential recommendation,” in2018 IEEE international conference on data mining (ICDM). IEEE, 2018, pp. 197–206. 1, 5, 6, 11","work_id":"726c788a-79e8-4201-834b-be022e969ccc","year":2018},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Contrastive learning for sequential recommendation,","work_id":"5ab05224-583d-4f76-bc70-9fd87386a327","year":2022}],"snapshot_sha256":"b1c6097f1cf2b944a00e3f57152c8f5708f2dc2489b057253ddfb6173415f623"},"source":{"id":"2605.17267","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T23:16:31.126607Z","id":"5939590b-8fab-4eba-b0eb-975c19f96497","model_set":{"reader":"grok-4.3"},"one_line_summary":"RAGR builds mixed item-review sequences for generative recommendation and uses DPO alignment to favor item tokens, reporting gains over prior GR baselines on three datasets.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Interleaving review semantic IDs into item sequences lets generative models use explanatory feedback to improve next-item predictions.","strongest_claim":"Experiments on three real-world datasets show that RAGR yields consistent and significant gains over strong GR backbones across all metrics.","weakest_assumption":"That interleaving review semantic IDs into the generative user sequence allows review signals to participate directly in autoregressive next-token generation while the Item-Centric Task Generation Alignment based on DPO successfully preserves the pure recommendation objective without introducing new biases or degrading item prediction quality."}},"verdict_id":"5939590b-8fab-4eba-b0eb-975c19f96497"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:a7179449eaff191ad9afebf21a30847a4cfecfdafb27b6facc56a9cf50ad8bed","target":"record","created_at":"2026-05-20T00:03:48Z","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":"beaed6b8b70a1ce6254095d23e067deb0778ddae9ebc246ae260e53d6be7db56","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2026-05-17T05:21:23Z","title_canon_sha256":"0da7fb68f4cf3afead675656b05cbbde0f33099287aaa483aea79ec6c3adb077"},"schema_version":"1.0","source":{"id":"2605.17267","kind":"arxiv","version":1}},"canonical_sha256":"8060d652d8a14c599bf796b09e48bb26b077e3845c508e042f8c6d7596a45385","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8060d652d8a14c599bf796b09e48bb26b077e3845c508e042f8c6d7596a45385","first_computed_at":"2026-05-20T00:03:48.864260Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:03:48.864260Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"fFpw69ywGIhPnPVUOsKeVTruMQ9FtJNC8pPQoCxzE6pEU79NKrg1L3BNDk5N91wswfR8SU7ovFE1SHwLIdkaBQ==","signature_status":"signed_v1","signed_at":"2026-05-20T00:03:48.865153Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.17267","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a7179449eaff191ad9afebf21a30847a4cfecfdafb27b6facc56a9cf50ad8bed","sha256:ad95ec8a344dda252cba6c27c632f2c20b317f2527df9e72e3e9da61298e973e"],"state_sha256":"8adcc8c43347868c4c0fc00bd693c89d4fc6c21cba001022900f73642fb23f87"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rDXE5baMGT+Z8VjJI5HzCeY/8m2ZNTUGdKPFzPNa38Ty71tvvxF8U1GSb/DirkAGMxZwHMN94svR3Bny02DgCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-29T18:00:12.156744Z","bundle_sha256":"485aa736284fd6b549d92af762e06e6201e278dd767ef98c3597654af05a910e"}}