{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:RQ4AQ6KGRT4TBOY3HURYNLFMP6","short_pith_number":"pith:RQ4AQ6KG","canonical_record":{"source":{"id":"2410.21662","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T02:11:45Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f03d4edda1cbddb55b3e41b15544148895cedc74278896c63a24b02b8461dd0b","abstract_canon_sha256":"6f0855a0f907c265fb2c596a60915e060084211015d1487f643d7b643b28e714"},"schema_version":"1.0"},"canonical_sha256":"8c380879468cf930bb1b3d2386acac7faf37e97ba9b8a8bc55123e59243448eb","source":{"kind":"arxiv","id":"2410.21662","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.21662","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"arxiv_version","alias_value":"2410.21662v2","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.21662","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_12","alias_value":"RQ4AQ6KGRT4T","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_16","alias_value":"RQ4AQ6KGRT4TBOY3","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_8","alias_value":"RQ4AQ6KG","created_at":"2026-07-05T10:14:47Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:RQ4AQ6KGRT4TBOY3HURYNLFMP6","target":"record","payload":{"canonical_record":{"source":{"id":"2410.21662","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T02:11:45Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f03d4edda1cbddb55b3e41b15544148895cedc74278896c63a24b02b8461dd0b","abstract_canon_sha256":"6f0855a0f907c265fb2c596a60915e060084211015d1487f643d7b643b28e714"},"schema_version":"1.0"},"canonical_sha256":"8c380879468cf930bb1b3d2386acac7faf37e97ba9b8a8bc55123e59243448eb","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:14:47.359858Z","signature_b64":"Im/B7AfclyKNXat8zoS5q/+p+bBwfuTMFQJKwhN/M9shVzCRu1F9eqLLaHDTP5Yw2gnKxfpGDivUF4cAxyyIBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8c380879468cf930bb1b3d2386acac7faf37e97ba9b8a8bc55123e59243448eb","last_reissued_at":"2026-07-05T10:14:47.359192Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:14:47.359192Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.21662","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:14:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"W4HOoP7Pm4om8RErXVRLbuAZlxIOsd3aHzhy3sWZZjv56xJBOek6RQckhk/EZ/pVcZ2K6z4DIbVdoNauD5tZBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T10:22:45.853781Z"},"content_sha256":"d05e549132b37fe5fcc70f586697c39f564305c410e0a36897954a9532499fd0","schema_version":"1.0","event_id":"sha256:d05e549132b37fe5fcc70f586697c39f564305c410e0a36897954a9532499fd0"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:RQ4AQ6KGRT4TBOY3HURYNLFMP6","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"$f$-PO: Generalizing Preference Optimization with $f$-divergence Minimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Jiaqi Han, Mingjian Jiang, Minkai Xu, Stefano Ermon, Yuxuan Song","submitted_at":"2024-10-29T02:11:45Z","abstract_excerpt":"Preference optimization has made significant progress recently, with numerous methods developed to align language models with human preferences. This paper introduces $f$-divergence Preference Optimization ($f$-PO), a novel framework that generalizes and extends existing approaches. $f$-PO minimizes $f$-divergences between the optimized policy and the optimal policy, encompassing a broad family of alignment methods using various divergences. Our approach unifies previous algorithms like DPO and EXO, while offering new variants through different choices of $f$-divergences. We provide theoretica"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.21662","kind":"arxiv","version":2},"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/2410.21662/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":"2026-07-05T10:14:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dyAj3SuGLLBj3OzX4oKAHh79Sv78kt1t5VqbMq+knO7gAwoql7vbxKa6AeUAmAJM0/MwL4qrts3+EXHoHW7iBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T10:22:45.854253Z"},"content_sha256":"c926615c4ee44641612b88bb577ddbc62ae532af3e4a8f50a4c9959aeadd145d","schema_version":"1.0","event_id":"sha256:c926615c4ee44641612b88bb577ddbc62ae532af3e4a8f50a4c9959aeadd145d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/RQ4AQ6KGRT4TBOY3HURYNLFMP6/bundle.json","state_url":"https://pith.science/pith/RQ4AQ6KGRT4TBOY3HURYNLFMP6/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/RQ4AQ6KGRT4TBOY3HURYNLFMP6/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":"2026-08-04T10:22:45Z","links":{"resolver":"https://pith.science/pith/RQ4AQ6KGRT4TBOY3HURYNLFMP6","bundle":"https://pith.science/pith/RQ4AQ6KGRT4TBOY3HURYNLFMP6/bundle.json","state":"https://pith.science/pith/RQ4AQ6KGRT4TBOY3HURYNLFMP6/state.json","well_known_bundle":"https://pith.science/.well-known/pith/RQ4AQ6KGRT4TBOY3HURYNLFMP6/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:RQ4AQ6KGRT4TBOY3HURYNLFMP6","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":"6f0855a0f907c265fb2c596a60915e060084211015d1487f643d7b643b28e714","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T02:11:45Z","title_canon_sha256":"f03d4edda1cbddb55b3e41b15544148895cedc74278896c63a24b02b8461dd0b"},"schema_version":"1.0","source":{"id":"2410.21662","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.21662","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"arxiv_version","alias_value":"2410.21662v2","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.21662","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_12","alias_value":"RQ4AQ6KGRT4T","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_16","alias_value":"RQ4AQ6KGRT4TBOY3","created_at":"2026-07-05T10:14:47Z"},{"alias_kind":"pith_short_8","alias_value":"RQ4AQ6KG","created_at":"2026-07-05T10:14:47Z"}],"graph_snapshots":[{"event_id":"sha256:c926615c4ee44641612b88bb577ddbc62ae532af3e4a8f50a4c9959aeadd145d","target":"graph","created_at":"2026-07-05T10:14:47Z","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/2410.21662/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Preference optimization has made significant progress recently, with numerous methods developed to align language models with human preferences. This paper introduces $f$-divergence Preference Optimization ($f$-PO), a novel framework that generalizes and extends existing approaches. $f$-PO minimizes $f$-divergences between the optimized policy and the optimal policy, encompassing a broad family of alignment methods using various divergences. Our approach unifies previous algorithms like DPO and EXO, while offering new variants through different choices of $f$-divergences. We provide theoretica","authors_text":"Jiaqi Han, Mingjian Jiang, Minkai Xu, Stefano Ermon, Yuxuan Song","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T02:11:45Z","title":"$f$-PO: Generalizing Preference Optimization with $f$-divergence Minimization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.21662","kind":"arxiv","version":2},"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:d05e549132b37fe5fcc70f586697c39f564305c410e0a36897954a9532499fd0","target":"record","created_at":"2026-07-05T10:14:47Z","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":"6f0855a0f907c265fb2c596a60915e060084211015d1487f643d7b643b28e714","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T02:11:45Z","title_canon_sha256":"f03d4edda1cbddb55b3e41b15544148895cedc74278896c63a24b02b8461dd0b"},"schema_version":"1.0","source":{"id":"2410.21662","kind":"arxiv","version":2}},"canonical_sha256":"8c380879468cf930bb1b3d2386acac7faf37e97ba9b8a8bc55123e59243448eb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8c380879468cf930bb1b3d2386acac7faf37e97ba9b8a8bc55123e59243448eb","first_computed_at":"2026-07-05T10:14:47.359192Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:14:47.359192Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Im/B7AfclyKNXat8zoS5q/+p+bBwfuTMFQJKwhN/M9shVzCRu1F9eqLLaHDTP5Yw2gnKxfpGDivUF4cAxyyIBQ==","signature_status":"signed_v1","signed_at":"2026-07-05T10:14:47.359858Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.21662","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d05e549132b37fe5fcc70f586697c39f564305c410e0a36897954a9532499fd0","sha256:c926615c4ee44641612b88bb577ddbc62ae532af3e4a8f50a4c9959aeadd145d"],"state_sha256":"714c43d337b05e7b5cb6dae3833d1a38c27b5407a2174d9718e9291d90e71bff"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"mzfX0Vn2/dEC1smm7XElRpPWDqbKdBVwkrow9UhzAVoqd0Xzd0rZxaMvM8pZ4kgpTi6wlr7NDve/tYQekSnJBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T10:22:45.859468Z","bundle_sha256":"3e998472ba50458d8ec2967c31463efcc461f0c91f10a6d505f0e952ce9068be"}}