{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:RCZODCXTBLSP3MJAXC4523AZHZ","short_pith_number":"pith:RCZODCXT","canonical_record":{"source":{"id":"2306.17492","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-06-30T09:07:37Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9466bf07c286bce839bb4794d0c33cd17dfe6b470b9eed6285a19786c7e5081f","abstract_canon_sha256":"70669fff3c425713c50060060e978dffc3c3c6e983fc4160892887114d6e2e40"},"schema_version":"1.0"},"canonical_sha256":"88b2e18af30ae4fdb120b8b9dd6c193e58e890f058a3742c84cd1f9bc16b0bbd","source":{"kind":"arxiv","id":"2306.17492","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2306.17492","created_at":"2026-07-05T07:49:55Z"},{"alias_kind":"arxiv_version","alias_value":"2306.17492v2","created_at":"2026-07-05T07:49:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2306.17492","created_at":"2026-07-05T07:49:55Z"},{"alias_kind":"pith_short_12","alias_value":"RCZODCXTBLSP","created_at":"2026-07-05T07:49:55Z"},{"alias_kind":"pith_short_16","alias_value":"RCZODCXTBLSP3MJA","created_at":"2026-07-05T07:49:55Z"},{"alias_kind":"pith_short_8","alias_value":"RCZODCXT","created_at":"2026-07-05T07:49:55Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:RCZODCXTBLSP3MJAXC4523AZHZ","target":"record","payload":{"canonical_record":{"source":{"id":"2306.17492","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-06-30T09:07:37Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9466bf07c286bce839bb4794d0c33cd17dfe6b470b9eed6285a19786c7e5081f","abstract_canon_sha256":"70669fff3c425713c50060060e978dffc3c3c6e983fc4160892887114d6e2e40"},"schema_version":"1.0"},"canonical_sha256":"88b2e18af30ae4fdb120b8b9dd6c193e58e890f058a3742c84cd1f9bc16b0bbd","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:49:55.230234Z","signature_b64":"nOCedPmvIN8Xnbz35noOa7O+0taVLdQSDMjk7miCBxhdQ6PIs4yDMqmDDZ8UCJR5b1KAKnSHzCG33SLYOgI6Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"88b2e18af30ae4fdb120b8b9dd6c193e58e890f058a3742c84cd1f9bc16b0bbd","last_reissued_at":"2026-07-05T07:49:55.229771Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:49:55.229771Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2306.17492","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":"2026-07-05T07:49:55Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Swkc413AlJ3EMj6ukA5sikHviiJqs4dhUPKZQz4Z/HnIU+l+wC1R1jVVwZC6XRzvgwJdUjKlECLyavUr7p/QCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T20:13:53.716386Z"},"content_sha256":"5eb99d288aec0b0fa3fe5258947b18a70b3f9ef1f0b0002367a13b4dd394a578","schema_version":"1.0","event_id":"sha256:5eb99d288aec0b0fa3fe5258947b18a70b3f9ef1f0b0002367a13b4dd394a578"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:RCZODCXTBLSP3MJAXC4523AZHZ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Preference Ranking Optimization for Human Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Bowen Yu, Feifan Song, Fei Huang, Haiyang Yu, Houfeng Wang, Minghao Li, Yongbin Li","submitted_at":"2023-06-30T09:07:37Z","abstract_excerpt":"Large language models (LLMs) often contain misleading content, emphasizing the need to align them with human values to ensure secure AI systems. Reinforcement learning from human feedback (RLHF) has been employed to achieve this alignment. However, it encompasses two main drawbacks: (1) RLHF exhibits complexity, instability, and sensitivity to hyperparameters in contrast to SFT. (2) Despite massive trial-and-error, multiple sampling is reduced to pair-wise contrast, thus lacking contrasts from a macro perspective. In this paper, we propose Preference Ranking Optimization (PRO) as an efficient "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2306.17492","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/2306.17492/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":"2026-07-05T07:49:55Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Qi17jPGdK16IF9b5G+71sK7BCXbDBf2NJTNuJfT57wt97DkdtG2Ej5197W8WaSimiyKakxd56ER/rhZy9g0MDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T20:13:53.717327Z"},"content_sha256":"619a8b2d08ad4b8b760060562f77816d1a59e82d2f246d30418076c9f7e05281","schema_version":"1.0","event_id":"sha256:619a8b2d08ad4b8b760060562f77816d1a59e82d2f246d30418076c9f7e05281"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/RCZODCXTBLSP3MJAXC4523AZHZ/bundle.json","state_url":"https://pith.science/pith/RCZODCXTBLSP3MJAXC4523AZHZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/RCZODCXTBLSP3MJAXC4523AZHZ/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":"2026-08-05T20:13:53Z","links":{"resolver":"https://pith.science/pith/RCZODCXTBLSP3MJAXC4523AZHZ","bundle":"https://pith.science/pith/RCZODCXTBLSP3MJAXC4523AZHZ/bundle.json","state":"https://pith.science/pith/RCZODCXTBLSP3MJAXC4523AZHZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/RCZODCXTBLSP3MJAXC4523AZHZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:RCZODCXTBLSP3MJAXC4523AZHZ","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":"70669fff3c425713c50060060e978dffc3c3c6e983fc4160892887114d6e2e40","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-06-30T09:07:37Z","title_canon_sha256":"9466bf07c286bce839bb4794d0c33cd17dfe6b470b9eed6285a19786c7e5081f"},"schema_version":"1.0","source":{"id":"2306.17492","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2306.17492","created_at":"2026-07-05T07:49:55Z"},{"alias_kind":"arxiv_version","alias_value":"2306.17492v2","created_at":"2026-07-05T07:49:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2306.17492","created_at":"2026-07-05T07:49:55Z"},{"alias_kind":"pith_short_12","alias_value":"RCZODCXTBLSP","created_at":"2026-07-05T07:49:55Z"},{"alias_kind":"pith_short_16","alias_value":"RCZODCXTBLSP3MJA","created_at":"2026-07-05T07:49:55Z"},{"alias_kind":"pith_short_8","alias_value":"RCZODCXT","created_at":"2026-07-05T07:49:55Z"}],"graph_snapshots":[{"event_id":"sha256:619a8b2d08ad4b8b760060562f77816d1a59e82d2f246d30418076c9f7e05281","target":"graph","created_at":"2026-07-05T07:49:55Z","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/2306.17492/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) often contain misleading content, emphasizing the need to align them with human values to ensure secure AI systems. Reinforcement learning from human feedback (RLHF) has been employed to achieve this alignment. However, it encompasses two main drawbacks: (1) RLHF exhibits complexity, instability, and sensitivity to hyperparameters in contrast to SFT. (2) Despite massive trial-and-error, multiple sampling is reduced to pair-wise contrast, thus lacking contrasts from a macro perspective. In this paper, we propose Preference Ranking Optimization (PRO) as an efficient ","authors_text":"Bowen Yu, Feifan Song, Fei Huang, Haiyang Yu, Houfeng Wang, Minghao Li, Yongbin Li","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-06-30T09:07:37Z","title":"Preference Ranking Optimization for Human Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2306.17492","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":"sha256:5eb99d288aec0b0fa3fe5258947b18a70b3f9ef1f0b0002367a13b4dd394a578","target":"record","created_at":"2026-07-05T07:49:55Z","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":"70669fff3c425713c50060060e978dffc3c3c6e983fc4160892887114d6e2e40","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-06-30T09:07:37Z","title_canon_sha256":"9466bf07c286bce839bb4794d0c33cd17dfe6b470b9eed6285a19786c7e5081f"},"schema_version":"1.0","source":{"id":"2306.17492","kind":"arxiv","version":2}},"canonical_sha256":"88b2e18af30ae4fdb120b8b9dd6c193e58e890f058a3742c84cd1f9bc16b0bbd","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"88b2e18af30ae4fdb120b8b9dd6c193e58e890f058a3742c84cd1f9bc16b0bbd","first_computed_at":"2026-07-05T07:49:55.229771Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:49:55.229771Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"nOCedPmvIN8Xnbz35noOa7O+0taVLdQSDMjk7miCBxhdQ6PIs4yDMqmDDZ8UCJR5b1KAKnSHzCG33SLYOgI6Dg==","signature_status":"signed_v1","signed_at":"2026-07-05T07:49:55.230234Z","signed_message":"canonical_sha256_bytes"},"source_id":"2306.17492","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:5eb99d288aec0b0fa3fe5258947b18a70b3f9ef1f0b0002367a13b4dd394a578","sha256:619a8b2d08ad4b8b760060562f77816d1a59e82d2f246d30418076c9f7e05281"],"state_sha256":"fc1cd8ab839993409119ba2c54aed9dda283890a6e2676024cf87d6b4df4f6ee"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1Hj2YeS4C/LjcPH9mRmZytG/RC1DpSmDN1T0wXct2LtuFI+Y2KIDVIrXmlggRDV+qppCd3ttOuTk7x8qsXgkCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T20:13:53.726102Z","bundle_sha256":"8aca12f37f41357b5bb566f4e39494c4c5f8d74c3c915877cb3f6629ea7033e0"}}