{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:2VX4IERKHRE2QEPTRB24XOZUH3","short_pith_number":"pith:2VX4IERK","canonical_record":{"source":{"id":"2505.22172","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-28T09:44:27Z","cross_cats_sorted":[],"title_canon_sha256":"6a2f7ad8feff08d8d94cbb940779a890207dd5f886580aa805dac2e9338dae08","abstract_canon_sha256":"4bd63fa9810c4110315bbc8fa81728217b64a7930e5596b024510973f7a2b025"},"schema_version":"1.0"},"canonical_sha256":"d56fc4122a3c49a811f38875cbbb343ee0499641b513216b85526170cffbf47c","source":{"kind":"arxiv","id":"2505.22172","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.22172","created_at":"2026-07-05T11:11:10Z"},{"alias_kind":"arxiv_version","alias_value":"2505.22172v1","created_at":"2026-07-05T11:11:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.22172","created_at":"2026-07-05T11:11:10Z"},{"alias_kind":"pith_short_12","alias_value":"2VX4IERKHRE2","created_at":"2026-07-05T11:11:10Z"},{"alias_kind":"pith_short_16","alias_value":"2VX4IERKHRE2QEPT","created_at":"2026-07-05T11:11:10Z"},{"alias_kind":"pith_short_8","alias_value":"2VX4IERK","created_at":"2026-07-05T11:11:10Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:2VX4IERKHRE2QEPTRB24XOZUH3","target":"record","payload":{"canonical_record":{"source":{"id":"2505.22172","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-28T09:44:27Z","cross_cats_sorted":[],"title_canon_sha256":"6a2f7ad8feff08d8d94cbb940779a890207dd5f886580aa805dac2e9338dae08","abstract_canon_sha256":"4bd63fa9810c4110315bbc8fa81728217b64a7930e5596b024510973f7a2b025"},"schema_version":"1.0"},"canonical_sha256":"d56fc4122a3c49a811f38875cbbb343ee0499641b513216b85526170cffbf47c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:11:10.912902Z","signature_b64":"e+mX6FRNJV8ng88pvr0Id8mZvgKWgwdsOfIea/31oYpHX0qdF/f2HAL6fEf9ha7KWEUWCRSgPLY+2ID65k5+Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d56fc4122a3c49a811f38875cbbb343ee0499641b513216b85526170cffbf47c","last_reissued_at":"2026-07-05T11:11:10.912352Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:11:10.912352Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.22172","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":"2026-07-05T11:11:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4zK9SUcHOTsXV3dgKhpOxGHM9t0LI9vEwCo/AVqex04xzDSS8/2Yw15i0DnnqaHBBnadljjH11hRGBZ/lMl+Bw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T09:31:42.457580Z"},"content_sha256":"d399201334711f71cfd6494205b148f7e46ef4e8770e17dbae95976e34defb2a","schema_version":"1.0","event_id":"sha256:d399201334711f71cfd6494205b148f7e46ef4e8770e17dbae95976e34defb2a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:2VX4IERKHRE2QEPTRB24XOZUH3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Reverse Preference Optimization for Complex Instruction Following","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Fei Huang, Feiteng Fang, Hangyu Li, Ting-En Lin, Xiang Huang, Yongbin Li, Yuchuan Wu, Yuzhong Qu","submitted_at":"2025-05-28T09:44:27Z","abstract_excerpt":"Instruction following (IF) is a critical capability for large language models (LLMs). However, handling complex instructions with multiple constraints remains challenging. Previous methods typically select preference pairs based on the number of constraints they satisfy, introducing noise where chosen examples may fail to follow some constraints and rejected examples may excel in certain respects over the chosen ones. To address the challenge of aligning with multiple preferences, we propose a simple yet effective method called Reverse Preference Optimization (RPO). It mitigates noise in prefe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.22172","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/2505.22172/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":"2026-07-05T11:11:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"YlW8ybmJd8MMYG1ArXg/Ooba+dAKEkJbLxbVKBi6YN4aeBszE8DDphcKPwOvD8nwx7kDbeb6yIsDl7aSx3/6CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T09:31:42.458512Z"},"content_sha256":"2bdb2c81ede0c0480e6e93af52fa992f3d73e57d205f1e85cdc64dffc0c73874","schema_version":"1.0","event_id":"sha256:2bdb2c81ede0c0480e6e93af52fa992f3d73e57d205f1e85cdc64dffc0c73874"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/2VX4IERKHRE2QEPTRB24XOZUH3/bundle.json","state_url":"https://pith.science/pith/2VX4IERKHRE2QEPTRB24XOZUH3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/2VX4IERKHRE2QEPTRB24XOZUH3/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":"2026-08-07T09:31:42Z","links":{"resolver":"https://pith.science/pith/2VX4IERKHRE2QEPTRB24XOZUH3","bundle":"https://pith.science/pith/2VX4IERKHRE2QEPTRB24XOZUH3/bundle.json","state":"https://pith.science/pith/2VX4IERKHRE2QEPTRB24XOZUH3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/2VX4IERKHRE2QEPTRB24XOZUH3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:2VX4IERKHRE2QEPTRB24XOZUH3","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":"4bd63fa9810c4110315bbc8fa81728217b64a7930e5596b024510973f7a2b025","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-28T09:44:27Z","title_canon_sha256":"6a2f7ad8feff08d8d94cbb940779a890207dd5f886580aa805dac2e9338dae08"},"schema_version":"1.0","source":{"id":"2505.22172","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.22172","created_at":"2026-07-05T11:11:10Z"},{"alias_kind":"arxiv_version","alias_value":"2505.22172v1","created_at":"2026-07-05T11:11:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.22172","created_at":"2026-07-05T11:11:10Z"},{"alias_kind":"pith_short_12","alias_value":"2VX4IERKHRE2","created_at":"2026-07-05T11:11:10Z"},{"alias_kind":"pith_short_16","alias_value":"2VX4IERKHRE2QEPT","created_at":"2026-07-05T11:11:10Z"},{"alias_kind":"pith_short_8","alias_value":"2VX4IERK","created_at":"2026-07-05T11:11:10Z"}],"graph_snapshots":[{"event_id":"sha256:2bdb2c81ede0c0480e6e93af52fa992f3d73e57d205f1e85cdc64dffc0c73874","target":"graph","created_at":"2026-07-05T11:11:10Z","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/2505.22172/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Instruction following (IF) is a critical capability for large language models (LLMs). However, handling complex instructions with multiple constraints remains challenging. Previous methods typically select preference pairs based on the number of constraints they satisfy, introducing noise where chosen examples may fail to follow some constraints and rejected examples may excel in certain respects over the chosen ones. To address the challenge of aligning with multiple preferences, we propose a simple yet effective method called Reverse Preference Optimization (RPO). It mitigates noise in prefe","authors_text":"Fei Huang, Feiteng Fang, Hangyu Li, Ting-En Lin, Xiang Huang, Yongbin Li, Yuchuan Wu, Yuzhong Qu","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-28T09:44:27Z","title":"Reverse Preference Optimization for Complex Instruction Following"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.22172","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:d399201334711f71cfd6494205b148f7e46ef4e8770e17dbae95976e34defb2a","target":"record","created_at":"2026-07-05T11:11:10Z","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":"4bd63fa9810c4110315bbc8fa81728217b64a7930e5596b024510973f7a2b025","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-28T09:44:27Z","title_canon_sha256":"6a2f7ad8feff08d8d94cbb940779a890207dd5f886580aa805dac2e9338dae08"},"schema_version":"1.0","source":{"id":"2505.22172","kind":"arxiv","version":1}},"canonical_sha256":"d56fc4122a3c49a811f38875cbbb343ee0499641b513216b85526170cffbf47c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d56fc4122a3c49a811f38875cbbb343ee0499641b513216b85526170cffbf47c","first_computed_at":"2026-07-05T11:11:10.912352Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:11:10.912352Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"e+mX6FRNJV8ng88pvr0Id8mZvgKWgwdsOfIea/31oYpHX0qdF/f2HAL6fEf9ha7KWEUWCRSgPLY+2ID65k5+Ag==","signature_status":"signed_v1","signed_at":"2026-07-05T11:11:10.912902Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.22172","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d399201334711f71cfd6494205b148f7e46ef4e8770e17dbae95976e34defb2a","sha256:2bdb2c81ede0c0480e6e93af52fa992f3d73e57d205f1e85cdc64dffc0c73874"],"state_sha256":"3e72f1770b606d1f0f6c006d54dc9aed2981410b71044787683789e065cde389"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jjUUUBwUOEXTfvzeKx3JoJ15buIM6Cj3iHJHMRO79vVodfBcBXSq9itcRln2DYU+IsGET9XmzCOgxumqxnfYAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T09:31:42.518176Z","bundle_sha256":"f159502c3f20a31d493dd42c76e62ff66e5d7c936e6d8d7f749ca268ce3e5b34"}}