{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:FJI76VL3Q5IXEDEK3P6EIRTC7V","short_pith_number":"pith:FJI76VL3","canonical_record":{"source":{"id":"2503.06358","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-03-08T23:41:20Z","cross_cats_sorted":[],"title_canon_sha256":"a0947de65793b1a528872f664bbcf8aaff2febdcfafe8c60879ec6cfed1cedf7","abstract_canon_sha256":"cd7c3368fa178312db08a416d77c3e22c18824a058b3bfc30135787830f72c3f"},"schema_version":"1.0"},"canonical_sha256":"2a51ff557b8751720c8adbfc444662fd59d34a214e445b0baaec2f8e1819f018","source":{"kind":"arxiv","id":"2503.06358","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.06358","created_at":"2026-07-05T10:27:31Z"},{"alias_kind":"arxiv_version","alias_value":"2503.06358v1","created_at":"2026-07-05T10:27:31Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.06358","created_at":"2026-07-05T10:27:31Z"},{"alias_kind":"pith_short_12","alias_value":"FJI76VL3Q5IX","created_at":"2026-07-05T10:27:31Z"},{"alias_kind":"pith_short_16","alias_value":"FJI76VL3Q5IXEDEK","created_at":"2026-07-05T10:27:31Z"},{"alias_kind":"pith_short_8","alias_value":"FJI76VL3","created_at":"2026-07-05T10:27:31Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:FJI76VL3Q5IXEDEK3P6EIRTC7V","target":"record","payload":{"canonical_record":{"source":{"id":"2503.06358","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-03-08T23:41:20Z","cross_cats_sorted":[],"title_canon_sha256":"a0947de65793b1a528872f664bbcf8aaff2febdcfafe8c60879ec6cfed1cedf7","abstract_canon_sha256":"cd7c3368fa178312db08a416d77c3e22c18824a058b3bfc30135787830f72c3f"},"schema_version":"1.0"},"canonical_sha256":"2a51ff557b8751720c8adbfc444662fd59d34a214e445b0baaec2f8e1819f018","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:27:31.555768Z","signature_b64":"1DD+JmI+dDqkl9Gjve/yPaxz+WqpzMXGCoe7DlHXUjXnKSVQlsvYY95benqOM3BwAOk04ieSZBplhWlqUPF5CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2a51ff557b8751720c8adbfc444662fd59d34a214e445b0baaec2f8e1819f018","last_reissued_at":"2026-07-05T10:27:31.555085Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:27:31.555085Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2503.06358","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":"2026-07-05T10:27:31Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"fz3kopmaej7Z9DR03zhVMDfV615C4xOGSdoLYTaImaaFmfloB/WPgRk/r0lxXTV8q3SaG6ROIV1A0cQZZhiNBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T09:04:46.979648Z"},"content_sha256":"f712cbab951e1713166f1a5909b55b5bf605279d8027018f6da1c84d3fabccab","schema_version":"1.0","event_id":"sha256:f712cbab951e1713166f1a5909b55b5bf605279d8027018f6da1c84d3fabccab"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:FJI76VL3Q5IXEDEK3P6EIRTC7V","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Language Model Personalization via Reward Factorization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Aldo Pacchiano, Felix Faltings, Idan Shenfeld, Pulkit Agrawal","submitted_at":"2025-03-08T23:41:20Z","abstract_excerpt":"Modern large language models (LLMs) are optimized for human-aligned responses using Reinforcement Learning from Human Feedback (RLHF). However, existing RLHF approaches assume a universal preference model and fail to account for individual user preferences, limiting their effectiveness in personalized applications. We introduce a framework that extends RLHF to enable user personalization by leveraging the assumption that user preferences lie in a low-dimensional space. Instead of training a separate model per user, we represent user-specific rewards as a linear combination of base reward funct"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.06358","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/2503.06358/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":"2026-07-05T10:27:31Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ROO6W0l0jg+LoGsxxU/on73/9bMxbc7bAjsbPaR0UWw4MSjaXonxP28DD1rmCBoI5nzjyTJUDW/4H/5ahhWGBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T09:04:46.980198Z"},"content_sha256":"313c391102c69e1626929316a4fbfefa88077a5722d925c0fd1141a66d509e3e","schema_version":"1.0","event_id":"sha256:313c391102c69e1626929316a4fbfefa88077a5722d925c0fd1141a66d509e3e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/FJI76VL3Q5IXEDEK3P6EIRTC7V/bundle.json","state_url":"https://pith.science/pith/FJI76VL3Q5IXEDEK3P6EIRTC7V/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/FJI76VL3Q5IXEDEK3P6EIRTC7V/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":"2026-08-06T09:04:46Z","links":{"resolver":"https://pith.science/pith/FJI76VL3Q5IXEDEK3P6EIRTC7V","bundle":"https://pith.science/pith/FJI76VL3Q5IXEDEK3P6EIRTC7V/bundle.json","state":"https://pith.science/pith/FJI76VL3Q5IXEDEK3P6EIRTC7V/state.json","well_known_bundle":"https://pith.science/.well-known/pith/FJI76VL3Q5IXEDEK3P6EIRTC7V/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:FJI76VL3Q5IXEDEK3P6EIRTC7V","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":"cd7c3368fa178312db08a416d77c3e22c18824a058b3bfc30135787830f72c3f","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-03-08T23:41:20Z","title_canon_sha256":"a0947de65793b1a528872f664bbcf8aaff2febdcfafe8c60879ec6cfed1cedf7"},"schema_version":"1.0","source":{"id":"2503.06358","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.06358","created_at":"2026-07-05T10:27:31Z"},{"alias_kind":"arxiv_version","alias_value":"2503.06358v1","created_at":"2026-07-05T10:27:31Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.06358","created_at":"2026-07-05T10:27:31Z"},{"alias_kind":"pith_short_12","alias_value":"FJI76VL3Q5IX","created_at":"2026-07-05T10:27:31Z"},{"alias_kind":"pith_short_16","alias_value":"FJI76VL3Q5IXEDEK","created_at":"2026-07-05T10:27:31Z"},{"alias_kind":"pith_short_8","alias_value":"FJI76VL3","created_at":"2026-07-05T10:27:31Z"}],"graph_snapshots":[{"event_id":"sha256:313c391102c69e1626929316a4fbfefa88077a5722d925c0fd1141a66d509e3e","target":"graph","created_at":"2026-07-05T10:27:31Z","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/2503.06358/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Modern large language models (LLMs) are optimized for human-aligned responses using Reinforcement Learning from Human Feedback (RLHF). However, existing RLHF approaches assume a universal preference model and fail to account for individual user preferences, limiting their effectiveness in personalized applications. We introduce a framework that extends RLHF to enable user personalization by leveraging the assumption that user preferences lie in a low-dimensional space. Instead of training a separate model per user, we represent user-specific rewards as a linear combination of base reward funct","authors_text":"Aldo Pacchiano, Felix Faltings, Idan Shenfeld, Pulkit Agrawal","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-03-08T23:41:20Z","title":"Language Model Personalization via Reward Factorization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.06358","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:f712cbab951e1713166f1a5909b55b5bf605279d8027018f6da1c84d3fabccab","target":"record","created_at":"2026-07-05T10:27:31Z","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":"cd7c3368fa178312db08a416d77c3e22c18824a058b3bfc30135787830f72c3f","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-03-08T23:41:20Z","title_canon_sha256":"a0947de65793b1a528872f664bbcf8aaff2febdcfafe8c60879ec6cfed1cedf7"},"schema_version":"1.0","source":{"id":"2503.06358","kind":"arxiv","version":1}},"canonical_sha256":"2a51ff557b8751720c8adbfc444662fd59d34a214e445b0baaec2f8e1819f018","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2a51ff557b8751720c8adbfc444662fd59d34a214e445b0baaec2f8e1819f018","first_computed_at":"2026-07-05T10:27:31.555085Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:27:31.555085Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"1DD+JmI+dDqkl9Gjve/yPaxz+WqpzMXGCoe7DlHXUjXnKSVQlsvYY95benqOM3BwAOk04ieSZBplhWlqUPF5CQ==","signature_status":"signed_v1","signed_at":"2026-07-05T10:27:31.555768Z","signed_message":"canonical_sha256_bytes"},"source_id":"2503.06358","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f712cbab951e1713166f1a5909b55b5bf605279d8027018f6da1c84d3fabccab","sha256:313c391102c69e1626929316a4fbfefa88077a5722d925c0fd1141a66d509e3e"],"state_sha256":"72ed58b1bc4ecd8b7e66bff0e5cd5bc0443cf7fb417e7a645ca19a645d6c174b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xB1dFraKINt4XMcjntKEMn08D77dxzoULb+MRQMKhtKegZskqX2fce6xUcwtCHUgXGqqc6nBYn4lFba9bXH5Bg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T09:04:46.985878Z","bundle_sha256":"5b16c318e989060db8c65831badff52b81436b983512afc089ccefa745350078"}}