{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:DJERVDEORMFD7KIAACCXJJEJGV","short_pith_number":"pith:DJERVDEO","canonical_record":{"source":{"id":"2608.01556","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-08-03T00:25:52Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"2543a0727cbf6339d8d0d6abd4d6e0080d9852e3d5fd235e97f183d649298763","abstract_canon_sha256":"1843317de42795e4a1360dc59c57ab3b7e1ce55d5a1b58cffc0cc0b04f05f076"},"schema_version":"1.0"},"canonical_sha256":"1a491a8c8e8b0a3fa900008574a489354f6319a1464d1ccfa3229e69cbb96883","source":{"kind":"arxiv","id":"2608.01556","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.01556","created_at":"2026-08-04T02:05:58Z"},{"alias_kind":"arxiv_version","alias_value":"2608.01556v1","created_at":"2026-08-04T02:05:58Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.01556","created_at":"2026-08-04T02:05:58Z"},{"alias_kind":"pith_short_12","alias_value":"DJERVDEORMFD","created_at":"2026-08-04T02:05:58Z"},{"alias_kind":"pith_short_16","alias_value":"DJERVDEORMFD7KIA","created_at":"2026-08-04T02:05:58Z"},{"alias_kind":"pith_short_8","alias_value":"DJERVDEO","created_at":"2026-08-04T02:05:58Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:DJERVDEORMFD7KIAACCXJJEJGV","target":"record","payload":{"canonical_record":{"source":{"id":"2608.01556","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-08-03T00:25:52Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"2543a0727cbf6339d8d0d6abd4d6e0080d9852e3d5fd235e97f183d649298763","abstract_canon_sha256":"1843317de42795e4a1360dc59c57ab3b7e1ce55d5a1b58cffc0cc0b04f05f076"},"schema_version":"1.0"},"canonical_sha256":"1a491a8c8e8b0a3fa900008574a489354f6319a1464d1ccfa3229e69cbb96883","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-04T02:05:58.589385Z","signature_b64":"Jx8KPKUrYpUIwnQIqwVHrq9LzRlDyBg9CT3f7giPZQ8m4eFE51JRiBBKVl956KtduCsIv7g2Os+uwV+w+vU3BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1a491a8c8e8b0a3fa900008574a489354f6319a1464d1ccfa3229e69cbb96883","last_reissued_at":"2026-08-04T02:05:58.587813Z","signature_status":"signed_v1","first_computed_at":"2026-08-04T02:05:58.587813Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2608.01556","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":"2026-08-04T02:05:58Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CqS6/t8+y2zv+rWYw382hRbE0+PapzByaNz6av8u3D1e84Q7W709ZdRpQQTqfy+xzNW4x4TFcWhdLjMh+lCYBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T10:17:43.697310Z"},"content_sha256":"dff2b5ab1ef2676f7286ba8e45a8915c780ad1532093e7d565d1ddf1ba5fe240","schema_version":"1.0","event_id":"sha256:dff2b5ab1ef2676f7286ba8e45a8915c780ad1532093e7d565d1ddf1ba5fe240"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:DJERVDEORMFD7KIAACCXJJEJGV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Rethinking Personalized Reward Modeling for LLMs under Preference Heterogeneity via Group-Debiased Federated Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Boryeong Cho, Jihwan Oh, Seokhyun Chung, Seongyoon Kim, Se-Young Yun","submitted_at":"2026-08-03T00:25:52Z","abstract_excerpt":"Large language models are increasingly aligned to human preferences via reward modeling, but user preference data are sensitive and often cannot be centralized. Federated learning keeps such data local while learning a shared initial reward model, which is later personalized for each client through local fine-tuning. Because users often assign opposite labels to the same pair of responses, existing federated methods address preference heterogeneity by clustering similar clients and training one reward model per group, assuming that each group requires its own initialization. We show that this "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.01556","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/2608.01556/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":"2026-08-04T02:05:58Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"E2ruPFhRocuvytN67Uu5+wBhD5wOrcuJGTeGujbq/71PfJ6QA9S6Q/e5nz/AvUDOKXLrGDKLt2f0x54ydp7vDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T10:17:43.697850Z"},"content_sha256":"0a4d9837baa942ef785c4b8567dd7de654c95616ec930a0104fabdeb5c1942ae","schema_version":"1.0","event_id":"sha256:0a4d9837baa942ef785c4b8567dd7de654c95616ec930a0104fabdeb5c1942ae"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:DJERVDEORMFD7KIAACCXJJEJGV","target":"integrity","payload":{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.18653/v1/2024.acl-long.468) was visible in the surrounding text but could not be confirmed against doi.org as printed.","snippet":"Haoxiang Wang, Yong Lin, Wei Xiong, Rui Yang, Shizhe Diao, Shuang Qiu, Han Zhao, and Tong Zhang. 2024. Arithmetic Control of LLMs for Diverse User Preferences: Directional Preference Alignment with Multi-Objective Rewards. InProceedings of ","arxiv_id":"2608.01556","detector":"doi_compliance","evidence":{"ref_index":47,"verdict_class":"incontrovertible","resolved_title":null,"printed_excerpt":"10.18653/v1/2024.ac","reconstructed_doi":"10.18653/v1/2024.acl-long.468"},"severity":"advisory","ref_index":47,"audited_at":"2026-08-07T00:23:27.079530Z","event_type":"pith.integrity.v1","detected_doi":"10.18653/v1/2024.acl-long.468","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"recoverable_identifier","evidence_hash":"5aa683d2d842b7e84fa5227af6920d61ad01703be842178b42c4782e3d637065","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.1.0","detected_arxiv_id":null,"integrity_event_id":18425,"payload_sha256":"3ebf710ae20e8f3aa4b3bb1ca7047e08e58ae9902f030b97e04496f72ba4a762","signature_b64":"vikTy7UZ7pc1eIlJikdsZfHwY3mWuorOGwuk+exELaMPPi3QwoRZOpgF+3tQHLei2vyQHjUyfM633kpUX6zBAw==","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":"2026-08-07T00:23:37Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"5rAs0GrziwIPdk607cVIjXieHJz/8xGcqC0DJnPeGRvN+HxPO7VirOzjm3oKIPZBcdYrmkyvFw9zKwYyfmADBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T10:17:43.702850Z"},"content_sha256":"a91975df5d3622052484c7bc8ede54af995f0d4265287b786d9fb3d7690ea87e","schema_version":"1.0","event_id":"sha256:a91975df5d3622052484c7bc8ede54af995f0d4265287b786d9fb3d7690ea87e"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:DJERVDEORMFD7KIAACCXJJEJGV","target":"integrity","payload":{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.1007/978-3-031-19809-0_10) was visible in the surrounding text but could not be confirmed against doi.org as printed.","snippet":"Xin Dong, Sai Qian Zhang, Ang Li, and H.T. Kung. 2022. SphereFed: Hyper- spherical Federated Learning. InComputer Vision – ECCV 2022: 17th European Conference, Tel A viv, Israel, October 23–27, 2022, Proceedings, Part XXVI(Tel Aviv, Israel)","arxiv_id":"2608.01556","detector":"doi_compliance","evidence":{"ref_index":10,"verdict_class":"incontrovertible","resolved_title":null,"printed_excerpt":"10.1007/978-","reconstructed_doi":"10.1007/978-3-031-19809-0_10"},"severity":"advisory","ref_index":10,"audited_at":"2026-08-07T00:23:27.079530Z","event_type":"pith.integrity.v1","detected_doi":"10.1007/978-3-031-19809-0_10","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"recoverable_identifier","evidence_hash":"b2cc2331bb98a3230ec314c3381b0c861c97d9b3360b2b9234cf99e631342996","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.1.0","detected_arxiv_id":null,"integrity_event_id":18424,"payload_sha256":"d5e9d702dbd4c116418924541e14121a59c3fa846816dadbc86ff29387d3908c","signature_b64":"M5jAsq8NZ0dJP1Cy+AGBub+XAAw8a7IMRHydE19AYR5SKl5C7pJWLfd2oaq/CKN8anonEhDYPOjLhBTadUusDQ==","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":"2026-08-07T00:23:37Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VoPcUrETBuUcyq3uXMGaLb+dT+6UW0nVMwuC9q/fel692BkIyImAw7IaF4FijAiF8AF02YnpV7i4+ePcWJoRDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T10:17:43.703404Z"},"content_sha256":"7111c86fb026645a241085b7cc59e6742283f84708200ef39221a4fe245f4e44","schema_version":"1.0","event_id":"sha256:7111c86fb026645a241085b7cc59e6742283f84708200ef39221a4fe245f4e44"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DJERVDEORMFD7KIAACCXJJEJGV/bundle.json","state_url":"https://pith.science/pith/DJERVDEORMFD7KIAACCXJJEJGV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DJERVDEORMFD7KIAACCXJJEJGV/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":"2026-08-09T10:17:43Z","links":{"resolver":"https://pith.science/pith/DJERVDEORMFD7KIAACCXJJEJGV","bundle":"https://pith.science/pith/DJERVDEORMFD7KIAACCXJJEJGV/bundle.json","state":"https://pith.science/pith/DJERVDEORMFD7KIAACCXJJEJGV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DJERVDEORMFD7KIAACCXJJEJGV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:DJERVDEORMFD7KIAACCXJJEJGV","merge_version":"pith-open-graph-merge-v1","event_count":4,"valid_event_count":4,"invalid_event_count":0,"equivocation_count":1,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"1843317de42795e4a1360dc59c57ab3b7e1ce55d5a1b58cffc0cc0b04f05f076","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-08-03T00:25:52Z","title_canon_sha256":"2543a0727cbf6339d8d0d6abd4d6e0080d9852e3d5fd235e97f183d649298763"},"schema_version":"1.0","source":{"id":"2608.01556","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.01556","created_at":"2026-08-04T02:05:58Z"},{"alias_kind":"arxiv_version","alias_value":"2608.01556v1","created_at":"2026-08-04T02:05:58Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.01556","created_at":"2026-08-04T02:05:58Z"},{"alias_kind":"pith_short_12","alias_value":"DJERVDEORMFD","created_at":"2026-08-04T02:05:58Z"},{"alias_kind":"pith_short_16","alias_value":"DJERVDEORMFD7KIA","created_at":"2026-08-04T02:05:58Z"},{"alias_kind":"pith_short_8","alias_value":"DJERVDEO","created_at":"2026-08-04T02:05:58Z"}],"graph_snapshots":[{"event_id":"sha256:0a4d9837baa942ef785c4b8567dd7de654c95616ec930a0104fabdeb5c1942ae","target":"graph","created_at":"2026-08-04T02:05:58Z","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/2608.01556/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models are increasingly aligned to human preferences via reward modeling, but user preference data are sensitive and often cannot be centralized. Federated learning keeps such data local while learning a shared initial reward model, which is later personalized for each client through local fine-tuning. Because users often assign opposite labels to the same pair of responses, existing federated methods address preference heterogeneity by clustering similar clients and training one reward model per group, assuming that each group requires its own initialization. We show that this ","authors_text":"Boryeong Cho, Jihwan Oh, Seokhyun Chung, Seongyoon Kim, Se-Young Yun","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-08-03T00:25:52Z","title":"Rethinking Personalized Reward Modeling for LLMs under Preference Heterogeneity via Group-Debiased Federated Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.01556","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:dff2b5ab1ef2676f7286ba8e45a8915c780ad1532093e7d565d1ddf1ba5fe240","target":"record","created_at":"2026-08-04T02:05:58Z","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":"1843317de42795e4a1360dc59c57ab3b7e1ce55d5a1b58cffc0cc0b04f05f076","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-08-03T00:25:52Z","title_canon_sha256":"2543a0727cbf6339d8d0d6abd4d6e0080d9852e3d5fd235e97f183d649298763"},"schema_version":"1.0","source":{"id":"2608.01556","kind":"arxiv","version":1}},"canonical_sha256":"1a491a8c8e8b0a3fa900008574a489354f6319a1464d1ccfa3229e69cbb96883","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1a491a8c8e8b0a3fa900008574a489354f6319a1464d1ccfa3229e69cbb96883","first_computed_at":"2026-08-04T02:05:58.587813Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-08-04T02:05:58.587813Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Jx8KPKUrYpUIwnQIqwVHrq9LzRlDyBg9CT3f7giPZQ8m4eFE51JRiBBKVl956KtduCsIv7g2Os+uwV+w+vU3BQ==","signature_status":"signed_v1","signed_at":"2026-08-04T02:05:58.589385Z","signed_message":"canonical_sha256_bytes"},"source_id":"2608.01556","source_kind":"arxiv","source_version":1}}},"equivocations":[{"signer_id":"pith.science","event_type":"integrity_finding","target":"integrity","event_ids":["sha256:7111c86fb026645a241085b7cc59e6742283f84708200ef39221a4fe245f4e44","sha256:a91975df5d3622052484c7bc8ede54af995f0d4265287b786d9fb3d7690ea87e"]}],"invalid_events":[],"applied_event_ids":["sha256:dff2b5ab1ef2676f7286ba8e45a8915c780ad1532093e7d565d1ddf1ba5fe240","sha256:0a4d9837baa942ef785c4b8567dd7de654c95616ec930a0104fabdeb5c1942ae"],"state_sha256":"66d668b4465ee6dfd3381d20fc84f9a966cee30fe1eeb1ce58112ec11e16a8fc"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"lOGH4h9FQZ4ZtTc692ySS1A8CosBz/rY5xp5+oU2m80j4Nr0f6IlCkfMFU06A05rMiuehp70vpbi9aK5xNVkDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T10:17:43.706421Z","bundle_sha256":"b2e01a9628b3c164b27381ca70fbc061e76b3b820f4a667e659a25a026b730cd"}}