{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:DH2723PBDAZ35ZYYH4YFZTHBTY","short_pith_number":"pith:DH2723PB","canonical_record":{"source":{"id":"2410.19720","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-25T17:47:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3c4c9ec3bd3081e8cd01e566d445e246469cee60d0e50fc77a0886b194506c9f","abstract_canon_sha256":"eaec649f9a9306cbbffee20111a308e64590aca93aeb3e7c45c4e4b1b034bd00"},"schema_version":"1.0"},"canonical_sha256":"19f5fd6de11833bee7183f305ccce19e0a46f2b769588f1e0698a97619175883","source":{"kind":"arxiv","id":"2410.19720","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.19720","created_at":"2026-07-05T09:26:00Z"},{"alias_kind":"arxiv_version","alias_value":"2410.19720v1","created_at":"2026-07-05T09:26:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.19720","created_at":"2026-07-05T09:26:00Z"},{"alias_kind":"pith_short_12","alias_value":"DH2723PBDAZ3","created_at":"2026-07-05T09:26:00Z"},{"alias_kind":"pith_short_16","alias_value":"DH2723PBDAZ35ZYY","created_at":"2026-07-05T09:26:00Z"},{"alias_kind":"pith_short_8","alias_value":"DH2723PB","created_at":"2026-07-05T09:26:00Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:DH2723PBDAZ35ZYYH4YFZTHBTY","target":"record","payload":{"canonical_record":{"source":{"id":"2410.19720","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-25T17:47:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3c4c9ec3bd3081e8cd01e566d445e246469cee60d0e50fc77a0886b194506c9f","abstract_canon_sha256":"eaec649f9a9306cbbffee20111a308e64590aca93aeb3e7c45c4e4b1b034bd00"},"schema_version":"1.0"},"canonical_sha256":"19f5fd6de11833bee7183f305ccce19e0a46f2b769588f1e0698a97619175883","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:26:00.796946Z","signature_b64":"BRvboYphUCEfjYh349CvZq5l/t6DUzLes9UnYKmJ/i2Y5LwBzwLt4H78h6Onm2ICOFOKnnEnSfv4oC7eFuuZDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"19f5fd6de11833bee7183f305ccce19e0a46f2b769588f1e0698a97619175883","last_reissued_at":"2026-07-05T09:26:00.796385Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:26:00.796385Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.19720","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":"2026-07-05T09:26:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"or9j0y541lOWNobiXc+ufZ+IOVYUdT0HDENnVUl8/jahpmlNOZOOkxaGKf6QPh8dmWxjq1wZHVlQsG46psglDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T05:59:02.485341Z"},"content_sha256":"b2f8c7aa2227b7e17ee95a25b47dbe736c92fe2693739556655f1d6b6dee53dd","schema_version":"1.0","event_id":"sha256:b2f8c7aa2227b7e17ee95a25b47dbe736c92fe2693739556655f1d6b6dee53dd"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:DH2723PBDAZ35ZYYH4YFZTHBTY","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"2D-DPO: Scaling Direct Preference Optimization with 2-Dimensional Supervision","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Bo Zheng, Hangyu Guo, Hui Huang, Jiaheng Liu, Jihao Gu, Shilong Li, Weixun Wang, Wenbo Su, Xingyuan Bu, Yancheng He","submitted_at":"2024-10-25T17:47:35Z","abstract_excerpt":"Recent advancements in Direct Preference Optimization (DPO) have significantly enhanced the alignment of Large Language Models (LLMs) with human preferences, owing to its simplicity and effectiveness. However, existing methods typically optimize a scalar score or ranking reward, thereby overlooking the multi-dimensional nature of human preferences. In this work, we propose to extend the preference of DPO to two dimensions: segments and aspects. We first introduce a 2D supervision dataset called HelpSteer-2D. For the segment dimension, we divide the response into sentences and assign scores to "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.19720","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/2410.19720/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":"2026-07-05T09:26:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Ewob1pIo1/OIRbjP9BArzNRM7aod8ldC4gDCMWV3sUdajHHqXSsYrHmAsP61isJuM8I8IA/Gs4UAmo8zyFbjBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T05:59:02.486000Z"},"content_sha256":"048d778c6d1a6a51124f450872e34dcd79726dbba54c6c0da6b3f2b761916a11","schema_version":"1.0","event_id":"sha256:048d778c6d1a6a51124f450872e34dcd79726dbba54c6c0da6b3f2b761916a11"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DH2723PBDAZ35ZYYH4YFZTHBTY/bundle.json","state_url":"https://pith.science/pith/DH2723PBDAZ35ZYYH4YFZTHBTY/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DH2723PBDAZ35ZYYH4YFZTHBTY/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":"2026-08-19T05:59:02Z","links":{"resolver":"https://pith.science/pith/DH2723PBDAZ35ZYYH4YFZTHBTY","bundle":"https://pith.science/pith/DH2723PBDAZ35ZYYH4YFZTHBTY/bundle.json","state":"https://pith.science/pith/DH2723PBDAZ35ZYYH4YFZTHBTY/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DH2723PBDAZ35ZYYH4YFZTHBTY/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:DH2723PBDAZ35ZYYH4YFZTHBTY","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":"eaec649f9a9306cbbffee20111a308e64590aca93aeb3e7c45c4e4b1b034bd00","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-25T17:47:35Z","title_canon_sha256":"3c4c9ec3bd3081e8cd01e566d445e246469cee60d0e50fc77a0886b194506c9f"},"schema_version":"1.0","source":{"id":"2410.19720","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.19720","created_at":"2026-07-05T09:26:00Z"},{"alias_kind":"arxiv_version","alias_value":"2410.19720v1","created_at":"2026-07-05T09:26:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.19720","created_at":"2026-07-05T09:26:00Z"},{"alias_kind":"pith_short_12","alias_value":"DH2723PBDAZ3","created_at":"2026-07-05T09:26:00Z"},{"alias_kind":"pith_short_16","alias_value":"DH2723PBDAZ35ZYY","created_at":"2026-07-05T09:26:00Z"},{"alias_kind":"pith_short_8","alias_value":"DH2723PB","created_at":"2026-07-05T09:26:00Z"}],"graph_snapshots":[{"event_id":"sha256:048d778c6d1a6a51124f450872e34dcd79726dbba54c6c0da6b3f2b761916a11","target":"graph","created_at":"2026-07-05T09:26:00Z","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/2410.19720/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent advancements in Direct Preference Optimization (DPO) have significantly enhanced the alignment of Large Language Models (LLMs) with human preferences, owing to its simplicity and effectiveness. However, existing methods typically optimize a scalar score or ranking reward, thereby overlooking the multi-dimensional nature of human preferences. In this work, we propose to extend the preference of DPO to two dimensions: segments and aspects. We first introduce a 2D supervision dataset called HelpSteer-2D. For the segment dimension, we divide the response into sentences and assign scores to ","authors_text":"Bo Zheng, Hangyu Guo, Hui Huang, Jiaheng Liu, Jihao Gu, Shilong Li, Weixun Wang, Wenbo Su, Xingyuan Bu, Yancheng He","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-25T17:47:35Z","title":"2D-DPO: Scaling Direct Preference Optimization with 2-Dimensional Supervision"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.19720","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:b2f8c7aa2227b7e17ee95a25b47dbe736c92fe2693739556655f1d6b6dee53dd","target":"record","created_at":"2026-07-05T09:26:00Z","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":"eaec649f9a9306cbbffee20111a308e64590aca93aeb3e7c45c4e4b1b034bd00","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-25T17:47:35Z","title_canon_sha256":"3c4c9ec3bd3081e8cd01e566d445e246469cee60d0e50fc77a0886b194506c9f"},"schema_version":"1.0","source":{"id":"2410.19720","kind":"arxiv","version":1}},"canonical_sha256":"19f5fd6de11833bee7183f305ccce19e0a46f2b769588f1e0698a97619175883","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"19f5fd6de11833bee7183f305ccce19e0a46f2b769588f1e0698a97619175883","first_computed_at":"2026-07-05T09:26:00.796385Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:26:00.796385Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"BRvboYphUCEfjYh349CvZq5l/t6DUzLes9UnYKmJ/i2Y5LwBzwLt4H78h6Onm2ICOFOKnnEnSfv4oC7eFuuZDA==","signature_status":"signed_v1","signed_at":"2026-07-05T09:26:00.796946Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.19720","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b2f8c7aa2227b7e17ee95a25b47dbe736c92fe2693739556655f1d6b6dee53dd","sha256:048d778c6d1a6a51124f450872e34dcd79726dbba54c6c0da6b3f2b761916a11"],"state_sha256":"417b61be945e442f540a51d93766349e0022f0067f4bcf8795835c9dd8ffa3d2"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ikbkuCl35j/k5yM+gZBZq0sbwVh711sVzKiPz1HuwWs+NMlu99YKS2stJjmL2OVuPDPVJW6DGufIQOKRFMIsBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-19T05:59:02.491541Z","bundle_sha256":"219024bb6c75615d84abcd11907d95102e365d636a7a880390482c3bf1227be7"}}