{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:CGWZ3LAXF7HEEVZTVUME2VJX5W","short_pith_number":"pith:CGWZ3LAX","canonical_record":{"source":{"id":"2507.20181","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-27T08:55:40Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"cc5edb0fb812fb525ddf612a7cf5ff65908fd7c725a8d8fb771a33764adfe385","abstract_canon_sha256":"aa46464b21753ea94535f909c9c75cd8c545d09cf64031b7e53392385655dc74"},"schema_version":"1.0"},"canonical_sha256":"11ad9dac172fce425733ad184d5537ed888992d656831538d49cc92286ed4ef1","source":{"kind":"arxiv","id":"2507.20181","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.20181","created_at":"2026-07-05T11:44:14Z"},{"alias_kind":"arxiv_version","alias_value":"2507.20181v1","created_at":"2026-07-05T11:44:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.20181","created_at":"2026-07-05T11:44:14Z"},{"alias_kind":"pith_short_12","alias_value":"CGWZ3LAXF7HE","created_at":"2026-07-05T11:44:14Z"},{"alias_kind":"pith_short_16","alias_value":"CGWZ3LAXF7HEEVZT","created_at":"2026-07-05T11:44:14Z"},{"alias_kind":"pith_short_8","alias_value":"CGWZ3LAX","created_at":"2026-07-05T11:44:14Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:CGWZ3LAXF7HEEVZTVUME2VJX5W","target":"record","payload":{"canonical_record":{"source":{"id":"2507.20181","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-27T08:55:40Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"cc5edb0fb812fb525ddf612a7cf5ff65908fd7c725a8d8fb771a33764adfe385","abstract_canon_sha256":"aa46464b21753ea94535f909c9c75cd8c545d09cf64031b7e53392385655dc74"},"schema_version":"1.0"},"canonical_sha256":"11ad9dac172fce425733ad184d5537ed888992d656831538d49cc92286ed4ef1","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:44:14.783566Z","signature_b64":"P4CwJUQ1Q9oIUQYc/2Prs2VULTFDcJotHED2yOibJvBT2qrny3oWgFrOJn4J0/Z8UPmnBLm7M9YOvD0EKLCQDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"11ad9dac172fce425733ad184d5537ed888992d656831538d49cc92286ed4ef1","last_reissued_at":"2026-07-05T11:44:14.783092Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:44:14.783092Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2507.20181","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":"2026-07-05T11:44:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+7qlqAwi0bQx/78i/N2PXQptmX5s7KxigbAb3KhFk4Y1AG8dD/ZyT58+xEe0Y2d9K5tB7k2N1JYJlZEx8tzRCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T15:45:32.456477Z"},"content_sha256":"b46148aa5d06bc13e622d9a3fba0282fc95b38dda83def238334a8fe82d35bde","schema_version":"1.0","event_id":"sha256:b46148aa5d06bc13e622d9a3fba0282fc95b38dda83def238334a8fe82d35bde"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:CGWZ3LAXF7HEEVZTVUME2VJX5W","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"SGPO: Self-Generated Preference Optimization based on Self-Improver","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Daejin Jo, Hyeonji Lee, Seohwan Yun, Sungwoong Kim","submitted_at":"2025-07-27T08:55:40Z","abstract_excerpt":"Large language models (LLMs), despite their extensive pretraining on diverse datasets, require effective alignment to human preferences for practical and reliable deployment. Conventional alignment methods typically employ off-policy learning and depend on human-annotated datasets, which limits their broad applicability and introduces distribution shift issues during training. To address these challenges, we propose Self-Generated Preference Optimization based on Self-Improver (SGPO), an innovative alignment framework that leverages an on-policy self-improving mechanism. Specifically, the impr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.20181","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/2507.20181/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":"2026-07-05T11:44:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"F4Y4CNbcG9RavHcJEPbVWKk5afvodTUHpK9gqVT/5rGOxTp4RwS/O/O+dx2JIDeNxZ5t+A0155z5q3znhcRlBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T15:45:32.457475Z"},"content_sha256":"6c2681a13de1187b802154d09c2420c069f1f4cfa127628a45a7b66654a581bd","schema_version":"1.0","event_id":"sha256:6c2681a13de1187b802154d09c2420c069f1f4cfa127628a45a7b66654a581bd"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/CGWZ3LAXF7HEEVZTVUME2VJX5W/bundle.json","state_url":"https://pith.science/pith/CGWZ3LAXF7HEEVZTVUME2VJX5W/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/CGWZ3LAXF7HEEVZTVUME2VJX5W/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":"2026-08-09T15:45:32Z","links":{"resolver":"https://pith.science/pith/CGWZ3LAXF7HEEVZTVUME2VJX5W","bundle":"https://pith.science/pith/CGWZ3LAXF7HEEVZTVUME2VJX5W/bundle.json","state":"https://pith.science/pith/CGWZ3LAXF7HEEVZTVUME2VJX5W/state.json","well_known_bundle":"https://pith.science/.well-known/pith/CGWZ3LAXF7HEEVZTVUME2VJX5W/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:CGWZ3LAXF7HEEVZTVUME2VJX5W","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":"aa46464b21753ea94535f909c9c75cd8c545d09cf64031b7e53392385655dc74","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-27T08:55:40Z","title_canon_sha256":"cc5edb0fb812fb525ddf612a7cf5ff65908fd7c725a8d8fb771a33764adfe385"},"schema_version":"1.0","source":{"id":"2507.20181","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.20181","created_at":"2026-07-05T11:44:14Z"},{"alias_kind":"arxiv_version","alias_value":"2507.20181v1","created_at":"2026-07-05T11:44:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.20181","created_at":"2026-07-05T11:44:14Z"},{"alias_kind":"pith_short_12","alias_value":"CGWZ3LAXF7HE","created_at":"2026-07-05T11:44:14Z"},{"alias_kind":"pith_short_16","alias_value":"CGWZ3LAXF7HEEVZT","created_at":"2026-07-05T11:44:14Z"},{"alias_kind":"pith_short_8","alias_value":"CGWZ3LAX","created_at":"2026-07-05T11:44:14Z"}],"graph_snapshots":[{"event_id":"sha256:6c2681a13de1187b802154d09c2420c069f1f4cfa127628a45a7b66654a581bd","target":"graph","created_at":"2026-07-05T11:44:14Z","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/2507.20181/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs), despite their extensive pretraining on diverse datasets, require effective alignment to human preferences for practical and reliable deployment. Conventional alignment methods typically employ off-policy learning and depend on human-annotated datasets, which limits their broad applicability and introduces distribution shift issues during training. To address these challenges, we propose Self-Generated Preference Optimization based on Self-Improver (SGPO), an innovative alignment framework that leverages an on-policy self-improving mechanism. Specifically, the impr","authors_text":"Daejin Jo, Hyeonji Lee, Seohwan Yun, Sungwoong Kim","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-27T08:55:40Z","title":"SGPO: Self-Generated Preference Optimization based on Self-Improver"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.20181","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:b46148aa5d06bc13e622d9a3fba0282fc95b38dda83def238334a8fe82d35bde","target":"record","created_at":"2026-07-05T11:44:14Z","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":"aa46464b21753ea94535f909c9c75cd8c545d09cf64031b7e53392385655dc74","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-27T08:55:40Z","title_canon_sha256":"cc5edb0fb812fb525ddf612a7cf5ff65908fd7c725a8d8fb771a33764adfe385"},"schema_version":"1.0","source":{"id":"2507.20181","kind":"arxiv","version":1}},"canonical_sha256":"11ad9dac172fce425733ad184d5537ed888992d656831538d49cc92286ed4ef1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"11ad9dac172fce425733ad184d5537ed888992d656831538d49cc92286ed4ef1","first_computed_at":"2026-07-05T11:44:14.783092Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:44:14.783092Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"P4CwJUQ1Q9oIUQYc/2Prs2VULTFDcJotHED2yOibJvBT2qrny3oWgFrOJn4J0/Z8UPmnBLm7M9YOvD0EKLCQDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:44:14.783566Z","signed_message":"canonical_sha256_bytes"},"source_id":"2507.20181","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b46148aa5d06bc13e622d9a3fba0282fc95b38dda83def238334a8fe82d35bde","sha256:6c2681a13de1187b802154d09c2420c069f1f4cfa127628a45a7b66654a581bd"],"state_sha256":"3d2aa3df8b3137bf91c2382b91850aed73b9f0709a64454ad6a29e64b614ed0b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2u+sdrb5jERZI3UsU6AEViU1rNylAFXS+41lyrAQQ5Tu80Ijsg4Z4+9LdZqC5VOEFotwzrh7fcr1PjTIb+TnDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T15:45:32.479801Z","bundle_sha256":"ef76e0c69d19ad935ae4e90624c44c1f8fdd97208db11c36247fa1892c19c6ad"}}