{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:IIM72VJLNQY6UC5ATWAAXPTQ4H","short_pith_number":"pith:IIM72VJL","canonical_record":{"source":{"id":"2404.16792","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-25T17:39:50Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"782ac821f7b0d2e29620b0d8b3a6ea8928ca1dec6057ddd996513dcba4a4b5d0","abstract_canon_sha256":"520c8a13c51ef1be88926fac42eb3f191ef1626ced913f558d3f0562c3ae1984"},"schema_version":"1.0"},"canonical_sha256":"4219fd552b6c31ea0ba09d800bbe70e1cf76bd86edacb722fc45e423f43df4f3","source":{"kind":"arxiv","id":"2404.16792","version":5},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.16792","created_at":"2026-07-05T11:12:17Z"},{"alias_kind":"arxiv_version","alias_value":"2404.16792v5","created_at":"2026-07-05T11:12:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.16792","created_at":"2026-07-05T11:12:17Z"},{"alias_kind":"pith_short_12","alias_value":"IIM72VJLNQY6","created_at":"2026-07-05T11:12:17Z"},{"alias_kind":"pith_short_16","alias_value":"IIM72VJLNQY6UC5A","created_at":"2026-07-05T11:12:17Z"},{"alias_kind":"pith_short_8","alias_value":"IIM72VJL","created_at":"2026-07-05T11:12:17Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:IIM72VJLNQY6UC5ATWAAXPTQ4H","target":"record","payload":{"canonical_record":{"source":{"id":"2404.16792","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-25T17:39:50Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"782ac821f7b0d2e29620b0d8b3a6ea8928ca1dec6057ddd996513dcba4a4b5d0","abstract_canon_sha256":"520c8a13c51ef1be88926fac42eb3f191ef1626ced913f558d3f0562c3ae1984"},"schema_version":"1.0"},"canonical_sha256":"4219fd552b6c31ea0ba09d800bbe70e1cf76bd86edacb722fc45e423f43df4f3","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:12:17.261419Z","signature_b64":"/aP54Y0LU/XWNZqj8sytLI9XU/P6xGcGWNsaOQ0NXMj74KEnP9AqznJMk1Tu2fDzdbDgiA822FyE67gNwtqvDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4219fd552b6c31ea0ba09d800bbe70e1cf76bd86edacb722fc45e423f43df4f3","last_reissued_at":"2026-07-05T11:12:17.260801Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:12:17.260801Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2404.16792","source_version":5,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:12:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8uva8C+ZryzB8mIn32x83lurCtjzFJSHlRkib73c1vgnOMi1E0BXnGEqCQESAcw1RvixOp38SbYXSsc1lWuaDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-17T15:14:02.512857Z"},"content_sha256":"d4282a7dfc6d2df3b05af92ca4b82618b69786ce3bfc7ec8de132d94520bc1bb","schema_version":"1.0","event_id":"sha256:d4282a7dfc6d2df3b05af92ca4b82618b69786ce3bfc7ec8de132d94520bc1bb"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:IIM72VJLNQY6UC5ATWAAXPTQ4H","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Model Extrapolation Expedites Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Chujie Zheng, Heng Ji, Minlie Huang, Nanyun Peng, Ziqi Wang","submitted_at":"2024-04-25T17:39:50Z","abstract_excerpt":"Given the high computational cost of preference alignment training of large language models (LLMs), exploring efficient methods to reduce the training overhead remains an important and compelling research problem. Motivated by the observation that alignment training typically involves only small parameter changes without injecting new knowledge into models, we propose a straightforward method called ExPO (model extrapolation) to expedite LLMs' alignment with human preferences. Given a partially-trained model and its initial SFT checkpoint, ExPO improves the implicit optimization objective of a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.16792","kind":"arxiv","version":5},"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/2404.16792/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":"2026-07-05T11:12:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"GLaQXE18+DSatXfeh2zbr7WrMi9yIVAYYJtKdJhD1t60nx3iI1zOCTFbjnZ8R5TRkOH++pk3SmOWScc1NoQKCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-17T15:14:02.513249Z"},"content_sha256":"8e3bf722e8fdfbe5b074fbaba91c5730beb52d4a81bd0357e5db079aee040023","schema_version":"1.0","event_id":"sha256:8e3bf722e8fdfbe5b074fbaba91c5730beb52d4a81bd0357e5db079aee040023"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/IIM72VJLNQY6UC5ATWAAXPTQ4H/bundle.json","state_url":"https://pith.science/pith/IIM72VJLNQY6UC5ATWAAXPTQ4H/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/IIM72VJLNQY6UC5ATWAAXPTQ4H/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":"2026-08-17T15:14:02Z","links":{"resolver":"https://pith.science/pith/IIM72VJLNQY6UC5ATWAAXPTQ4H","bundle":"https://pith.science/pith/IIM72VJLNQY6UC5ATWAAXPTQ4H/bundle.json","state":"https://pith.science/pith/IIM72VJLNQY6UC5ATWAAXPTQ4H/state.json","well_known_bundle":"https://pith.science/.well-known/pith/IIM72VJLNQY6UC5ATWAAXPTQ4H/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:IIM72VJLNQY6UC5ATWAAXPTQ4H","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":"520c8a13c51ef1be88926fac42eb3f191ef1626ced913f558d3f0562c3ae1984","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-25T17:39:50Z","title_canon_sha256":"782ac821f7b0d2e29620b0d8b3a6ea8928ca1dec6057ddd996513dcba4a4b5d0"},"schema_version":"1.0","source":{"id":"2404.16792","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.16792","created_at":"2026-07-05T11:12:17Z"},{"alias_kind":"arxiv_version","alias_value":"2404.16792v5","created_at":"2026-07-05T11:12:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.16792","created_at":"2026-07-05T11:12:17Z"},{"alias_kind":"pith_short_12","alias_value":"IIM72VJLNQY6","created_at":"2026-07-05T11:12:17Z"},{"alias_kind":"pith_short_16","alias_value":"IIM72VJLNQY6UC5A","created_at":"2026-07-05T11:12:17Z"},{"alias_kind":"pith_short_8","alias_value":"IIM72VJL","created_at":"2026-07-05T11:12:17Z"}],"graph_snapshots":[{"event_id":"sha256:8e3bf722e8fdfbe5b074fbaba91c5730beb52d4a81bd0357e5db079aee040023","target":"graph","created_at":"2026-07-05T11:12:17Z","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/2404.16792/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Given the high computational cost of preference alignment training of large language models (LLMs), exploring efficient methods to reduce the training overhead remains an important and compelling research problem. Motivated by the observation that alignment training typically involves only small parameter changes without injecting new knowledge into models, we propose a straightforward method called ExPO (model extrapolation) to expedite LLMs' alignment with human preferences. Given a partially-trained model and its initial SFT checkpoint, ExPO improves the implicit optimization objective of a","authors_text":"Chujie Zheng, Heng Ji, Minlie Huang, Nanyun Peng, Ziqi Wang","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-25T17:39:50Z","title":"Model Extrapolation Expedites Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.16792","kind":"arxiv","version":5},"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:d4282a7dfc6d2df3b05af92ca4b82618b69786ce3bfc7ec8de132d94520bc1bb","target":"record","created_at":"2026-07-05T11:12:17Z","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":"520c8a13c51ef1be88926fac42eb3f191ef1626ced913f558d3f0562c3ae1984","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-25T17:39:50Z","title_canon_sha256":"782ac821f7b0d2e29620b0d8b3a6ea8928ca1dec6057ddd996513dcba4a4b5d0"},"schema_version":"1.0","source":{"id":"2404.16792","kind":"arxiv","version":5}},"canonical_sha256":"4219fd552b6c31ea0ba09d800bbe70e1cf76bd86edacb722fc45e423f43df4f3","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4219fd552b6c31ea0ba09d800bbe70e1cf76bd86edacb722fc45e423f43df4f3","first_computed_at":"2026-07-05T11:12:17.260801Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:12:17.260801Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"/aP54Y0LU/XWNZqj8sytLI9XU/P6xGcGWNsaOQ0NXMj74KEnP9AqznJMk1Tu2fDzdbDgiA822FyE67gNwtqvDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:12:17.261419Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.16792","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d4282a7dfc6d2df3b05af92ca4b82618b69786ce3bfc7ec8de132d94520bc1bb","sha256:8e3bf722e8fdfbe5b074fbaba91c5730beb52d4a81bd0357e5db079aee040023"],"state_sha256":"3d81a242e335b312689b15808e1c68e3fbe20a37447375bf17eeb4355144c766"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HcOvSmFDf1H5dp73pLcFXAmqftvPGSQRbXDh4fTfjThV29GBkVjp1xW9a++KA8MrNVu3ibBNb+lXDjpE/AtpAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-17T15:14:02.516798Z","bundle_sha256":"7c814cb90ab6b29e447e7543d46836d49a1be815fb7ebbd612ceb360f3b8c78b"}}