{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:EZYO4FZI5BTUFTQQHN2BC57NVJ","short_pith_number":"pith:EZYO4FZI","canonical_record":{"source":{"id":"2402.11253","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-17T11:25:26Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"dd72d02b5e3f8b3b81f4290d713a4a062b7dde79219bd5b96cc40ba385df43e2","abstract_canon_sha256":"9fb6b5ef821400383a940c5965c49a0497e99950c9681f9051259258c9216bde"},"schema_version":"1.0"},"canonical_sha256":"2670ee1728e86742ce103b741177edaa77fcc95396509e84ca9d980eb9b467da","source":{"kind":"arxiv","id":"2402.11253","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.11253","created_at":"2026-07-05T08:36:22Z"},{"alias_kind":"arxiv_version","alias_value":"2402.11253v3","created_at":"2026-07-05T08:36:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.11253","created_at":"2026-07-05T08:36:22Z"},{"alias_kind":"pith_short_12","alias_value":"EZYO4FZI5BTU","created_at":"2026-07-05T08:36:22Z"},{"alias_kind":"pith_short_16","alias_value":"EZYO4FZI5BTUFTQQ","created_at":"2026-07-05T08:36:22Z"},{"alias_kind":"pith_short_8","alias_value":"EZYO4FZI","created_at":"2026-07-05T08:36:22Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:EZYO4FZI5BTUFTQQHN2BC57NVJ","target":"record","payload":{"canonical_record":{"source":{"id":"2402.11253","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-17T11:25:26Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"dd72d02b5e3f8b3b81f4290d713a4a062b7dde79219bd5b96cc40ba385df43e2","abstract_canon_sha256":"9fb6b5ef821400383a940c5965c49a0497e99950c9681f9051259258c9216bde"},"schema_version":"1.0"},"canonical_sha256":"2670ee1728e86742ce103b741177edaa77fcc95396509e84ca9d980eb9b467da","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:36:22.676891Z","signature_b64":"AC2GgPfBWDbM5k9muiob4HGkfv4dMgEJrYEt2FcuTIaI9HJRESuJ/FhdrQbIPhy6eOs3KwuGrQ1BLX0YFE0kDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2670ee1728e86742ce103b741177edaa77fcc95396509e84ca9d980eb9b467da","last_reissued_at":"2026-07-05T08:36:22.676408Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:36:22.676408Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.11253","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:36:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1B9q9WNvusO8X5jp32lXLaUj5r0GMXBJppYgsSLrGcHqS023eyOgBG6EeDJm0m5CcYdeCV2FJLH1qPLcEuQwDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T10:55:28.238989Z"},"content_sha256":"363ae070f860e4452b0f98489bfd3aa4c83ccd08976adc1d2fc1fe579cd6cac2","schema_version":"1.0","event_id":"sha256:363ae070f860e4452b0f98489bfd3aa4c83ccd08976adc1d2fc1fe579cd6cac2"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:EZYO4FZI5BTUFTQQHN2BC57NVJ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Aligning Large Language Models by On-Policy Self-Judgment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Ashkan Yousefpour, Kang Min Yoo, Minjoon Seo, Sangkyu Lee, Sungdong Kim, Youngjae Yu","submitted_at":"2024-02-17T11:25:26Z","abstract_excerpt":"Existing approaches for aligning large language models with human preferences face a trade-off that requires a separate reward model (RM) for on-policy learning. In this paper, we present a novel alignment framework, SELF-JUDGE that (1) does on-policy learning and 2) is parameter efficient, as it does not require an additional RM for evaluating the samples for on-policy learning. To this end, we propose Judge-augmented Supervised Fine-Tuning (JSFT) to train a single model to act as both a policy and a judge. Specifically, we view the pairwise judgment task, choosing the better response from a "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.11253","kind":"arxiv","version":3},"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/2402.11253/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":"2026-07-05T08:36:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Z7YycRFkGIOANEIGH01WlzpTJAm3Oy151tmbDY69Oq7MhIzfrT8w0WWvmdB21cL2XQYyb8XJfEeHvJntUslIAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T10:55:28.239971Z"},"content_sha256":"6d78f47a6f7eaf7b2bd604cd0210b4705488aa22d46ba49ea8fb5a7b571eaca0","schema_version":"1.0","event_id":"sha256:6d78f47a6f7eaf7b2bd604cd0210b4705488aa22d46ba49ea8fb5a7b571eaca0"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/EZYO4FZI5BTUFTQQHN2BC57NVJ/bundle.json","state_url":"https://pith.science/pith/EZYO4FZI5BTUFTQQHN2BC57NVJ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/EZYO4FZI5BTUFTQQHN2BC57NVJ/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":"2026-08-10T10:55:28Z","links":{"resolver":"https://pith.science/pith/EZYO4FZI5BTUFTQQHN2BC57NVJ","bundle":"https://pith.science/pith/EZYO4FZI5BTUFTQQHN2BC57NVJ/bundle.json","state":"https://pith.science/pith/EZYO4FZI5BTUFTQQHN2BC57NVJ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/EZYO4FZI5BTUFTQQHN2BC57NVJ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:EZYO4FZI5BTUFTQQHN2BC57NVJ","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":"9fb6b5ef821400383a940c5965c49a0497e99950c9681f9051259258c9216bde","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-17T11:25:26Z","title_canon_sha256":"dd72d02b5e3f8b3b81f4290d713a4a062b7dde79219bd5b96cc40ba385df43e2"},"schema_version":"1.0","source":{"id":"2402.11253","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.11253","created_at":"2026-07-05T08:36:22Z"},{"alias_kind":"arxiv_version","alias_value":"2402.11253v3","created_at":"2026-07-05T08:36:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.11253","created_at":"2026-07-05T08:36:22Z"},{"alias_kind":"pith_short_12","alias_value":"EZYO4FZI5BTU","created_at":"2026-07-05T08:36:22Z"},{"alias_kind":"pith_short_16","alias_value":"EZYO4FZI5BTUFTQQ","created_at":"2026-07-05T08:36:22Z"},{"alias_kind":"pith_short_8","alias_value":"EZYO4FZI","created_at":"2026-07-05T08:36:22Z"}],"graph_snapshots":[{"event_id":"sha256:6d78f47a6f7eaf7b2bd604cd0210b4705488aa22d46ba49ea8fb5a7b571eaca0","target":"graph","created_at":"2026-07-05T08:36:22Z","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/2402.11253/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Existing approaches for aligning large language models with human preferences face a trade-off that requires a separate reward model (RM) for on-policy learning. In this paper, we present a novel alignment framework, SELF-JUDGE that (1) does on-policy learning and 2) is parameter efficient, as it does not require an additional RM for evaluating the samples for on-policy learning. To this end, we propose Judge-augmented Supervised Fine-Tuning (JSFT) to train a single model to act as both a policy and a judge. Specifically, we view the pairwise judgment task, choosing the better response from a ","authors_text":"Ashkan Yousefpour, Kang Min Yoo, Minjoon Seo, Sangkyu Lee, Sungdong Kim, Youngjae Yu","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-17T11:25:26Z","title":"Aligning Large Language Models by On-Policy Self-Judgment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.11253","kind":"arxiv","version":3},"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":"sha256:363ae070f860e4452b0f98489bfd3aa4c83ccd08976adc1d2fc1fe579cd6cac2","target":"record","created_at":"2026-07-05T08:36:22Z","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":"9fb6b5ef821400383a940c5965c49a0497e99950c9681f9051259258c9216bde","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-17T11:25:26Z","title_canon_sha256":"dd72d02b5e3f8b3b81f4290d713a4a062b7dde79219bd5b96cc40ba385df43e2"},"schema_version":"1.0","source":{"id":"2402.11253","kind":"arxiv","version":3}},"canonical_sha256":"2670ee1728e86742ce103b741177edaa77fcc95396509e84ca9d980eb9b467da","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2670ee1728e86742ce103b741177edaa77fcc95396509e84ca9d980eb9b467da","first_computed_at":"2026-07-05T08:36:22.676408Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:36:22.676408Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"AC2GgPfBWDbM5k9muiob4HGkfv4dMgEJrYEt2FcuTIaI9HJRESuJ/FhdrQbIPhy6eOs3KwuGrQ1BLX0YFE0kDA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:36:22.676891Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.11253","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:363ae070f860e4452b0f98489bfd3aa4c83ccd08976adc1d2fc1fe579cd6cac2","sha256:6d78f47a6f7eaf7b2bd604cd0210b4705488aa22d46ba49ea8fb5a7b571eaca0"],"state_sha256":"e6bb7522a0024defc037359858630785dfb8ffd61d84f58f4e5a0b5e0d4f5f9e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"M0Z/ye2qjscZeFMpBN07tgxamBKcgOGvmxOrq5/gBEcb8KeDTPzJLIwA25vm3vna8Z7m2+5zlDWmpEHCGxfTCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T10:55:28.246583Z","bundle_sha256":"3825ec98a25e8ebaea8d0b3d4818d302831f17c661bab0f440944594a4656189"}}