{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:GCLLTN6TSNAQJYADEOLFH53BLT","short_pith_number":"pith:GCLLTN6T","canonical_record":{"source":{"id":"2402.11907","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-19T07:46:40Z","cross_cats_sorted":[],"title_canon_sha256":"081715fb2c285a67dcef2303f4908bb05314f7db3c1a49e9d5a161bac284a405","abstract_canon_sha256":"11e293d0aedc7a523444704439307ea883888cd947bc255215f42d60d88cb575"},"schema_version":"1.0"},"canonical_sha256":"3096b9b7d3934104e003239653f7615cd0bfd546e8f587fa5dce5e1ed5126bfc","source":{"kind":"arxiv","id":"2402.11907","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.11907","created_at":"2026-07-05T08:55:33Z"},{"alias_kind":"arxiv_version","alias_value":"2402.11907v2","created_at":"2026-07-05T08:55:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.11907","created_at":"2026-07-05T08:55:33Z"},{"alias_kind":"pith_short_12","alias_value":"GCLLTN6TSNAQ","created_at":"2026-07-05T08:55:33Z"},{"alias_kind":"pith_short_16","alias_value":"GCLLTN6TSNAQJYAD","created_at":"2026-07-05T08:55:33Z"},{"alias_kind":"pith_short_8","alias_value":"GCLLTN6T","created_at":"2026-07-05T08:55:33Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:GCLLTN6TSNAQJYADEOLFH53BLT","target":"record","payload":{"canonical_record":{"source":{"id":"2402.11907","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-19T07:46:40Z","cross_cats_sorted":[],"title_canon_sha256":"081715fb2c285a67dcef2303f4908bb05314f7db3c1a49e9d5a161bac284a405","abstract_canon_sha256":"11e293d0aedc7a523444704439307ea883888cd947bc255215f42d60d88cb575"},"schema_version":"1.0"},"canonical_sha256":"3096b9b7d3934104e003239653f7615cd0bfd546e8f587fa5dce5e1ed5126bfc","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:55:33.410171Z","signature_b64":"KhBvt24szFlDEdKr5i8wCGs8X+9gxPJh+DlAkVizreJa/qUsUyRMdgvTXgttjEj43mrrGHAXndw9YCsLk+DqDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3096b9b7d3934104e003239653f7615cd0bfd546e8f587fa5dce5e1ed5126bfc","last_reissued_at":"2026-07-05T08:55:33.409613Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:55:33.409613Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.11907","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":"2026-07-05T08:55:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8DjWAL0Vp50UOA555iRYvAUJlc1RZEI1/T3hpAafxAWs/WU6MrIUnJAiDF2660+CyQLqqVNMnJYKgM73IHWtAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T19:17:17.750668Z"},"content_sha256":"4a416b2b8a53146e49a5d92bcf0e571ccd5859ed9ed312436d945d8e23a84e82","schema_version":"1.0","event_id":"sha256:4a416b2b8a53146e49a5d92bcf0e571ccd5859ed9ed312436d945d8e23a84e82"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:GCLLTN6TSNAQJYADEOLFH53BLT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Direct Large Language Model Alignment Through Self-Rewarding Contrastive Prompt Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aiwei Liu, Haoping Bai, Jiulong Shan, Lijie Wen, Meng Cao, Simon Wang, Xiang Kong, Zhiyun Lu","submitted_at":"2024-02-19T07:46:40Z","abstract_excerpt":"Aligning large language models (LLMs) with human expectations without human-annotated preference data is an important problem. In this paper, we propose a method to evaluate the response preference by using the output probabilities of response pairs under contrastive prompt pairs, which could achieve better performance on LLaMA2-7B and LLaMA2-13B compared to RLAIF. Based on this, we propose an automatic alignment method, Direct Large Model Alignment (DLMA). First, we use contrastive prompt pairs to automatically generate preference data. Then, we continue to evaluate the generated preference d"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.11907","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/2402.11907/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":"2026-07-05T08:55:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HLrpYdTADh+YWeKx/SNRlgzKNaQb7FqJMyaNjYuM3j4GLXXM1c0Hch4I34mVpIOT9ocpq4ZBA2ix3mp6yQ94DQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T19:17:17.751043Z"},"content_sha256":"aa89c3782ea569b10f15e27a6f9452da6d1c8e2ae2df24b3fdb0fbeb8894d0b0","schema_version":"1.0","event_id":"sha256:aa89c3782ea569b10f15e27a6f9452da6d1c8e2ae2df24b3fdb0fbeb8894d0b0"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/GCLLTN6TSNAQJYADEOLFH53BLT/bundle.json","state_url":"https://pith.science/pith/GCLLTN6TSNAQJYADEOLFH53BLT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/GCLLTN6TSNAQJYADEOLFH53BLT/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":"2026-07-06T19:17:17Z","links":{"resolver":"https://pith.science/pith/GCLLTN6TSNAQJYADEOLFH53BLT","bundle":"https://pith.science/pith/GCLLTN6TSNAQJYADEOLFH53BLT/bundle.json","state":"https://pith.science/pith/GCLLTN6TSNAQJYADEOLFH53BLT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/GCLLTN6TSNAQJYADEOLFH53BLT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:GCLLTN6TSNAQJYADEOLFH53BLT","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":"11e293d0aedc7a523444704439307ea883888cd947bc255215f42d60d88cb575","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-19T07:46:40Z","title_canon_sha256":"081715fb2c285a67dcef2303f4908bb05314f7db3c1a49e9d5a161bac284a405"},"schema_version":"1.0","source":{"id":"2402.11907","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.11907","created_at":"2026-07-05T08:55:33Z"},{"alias_kind":"arxiv_version","alias_value":"2402.11907v2","created_at":"2026-07-05T08:55:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.11907","created_at":"2026-07-05T08:55:33Z"},{"alias_kind":"pith_short_12","alias_value":"GCLLTN6TSNAQ","created_at":"2026-07-05T08:55:33Z"},{"alias_kind":"pith_short_16","alias_value":"GCLLTN6TSNAQJYAD","created_at":"2026-07-05T08:55:33Z"},{"alias_kind":"pith_short_8","alias_value":"GCLLTN6T","created_at":"2026-07-05T08:55:33Z"}],"graph_snapshots":[{"event_id":"sha256:aa89c3782ea569b10f15e27a6f9452da6d1c8e2ae2df24b3fdb0fbeb8894d0b0","target":"graph","created_at":"2026-07-05T08:55:33Z","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/2402.11907/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Aligning large language models (LLMs) with human expectations without human-annotated preference data is an important problem. In this paper, we propose a method to evaluate the response preference by using the output probabilities of response pairs under contrastive prompt pairs, which could achieve better performance on LLaMA2-7B and LLaMA2-13B compared to RLAIF. Based on this, we propose an automatic alignment method, Direct Large Model Alignment (DLMA). First, we use contrastive prompt pairs to automatically generate preference data. Then, we continue to evaluate the generated preference d","authors_text":"Aiwei Liu, Haoping Bai, Jiulong Shan, Lijie Wen, Meng Cao, Simon Wang, Xiang Kong, Zhiyun Lu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-19T07:46:40Z","title":"Direct Large Language Model Alignment Through Self-Rewarding Contrastive Prompt Distillation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.11907","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":"sha256:4a416b2b8a53146e49a5d92bcf0e571ccd5859ed9ed312436d945d8e23a84e82","target":"record","created_at":"2026-07-05T08:55:33Z","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":"11e293d0aedc7a523444704439307ea883888cd947bc255215f42d60d88cb575","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-19T07:46:40Z","title_canon_sha256":"081715fb2c285a67dcef2303f4908bb05314f7db3c1a49e9d5a161bac284a405"},"schema_version":"1.0","source":{"id":"2402.11907","kind":"arxiv","version":2}},"canonical_sha256":"3096b9b7d3934104e003239653f7615cd0bfd546e8f587fa5dce5e1ed5126bfc","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3096b9b7d3934104e003239653f7615cd0bfd546e8f587fa5dce5e1ed5126bfc","first_computed_at":"2026-07-05T08:55:33.409613Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:55:33.409613Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KhBvt24szFlDEdKr5i8wCGs8X+9gxPJh+DlAkVizreJa/qUsUyRMdgvTXgttjEj43mrrGHAXndw9YCsLk+DqDw==","signature_status":"signed_v1","signed_at":"2026-07-05T08:55:33.410171Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.11907","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4a416b2b8a53146e49a5d92bcf0e571ccd5859ed9ed312436d945d8e23a84e82","sha256:aa89c3782ea569b10f15e27a6f9452da6d1c8e2ae2df24b3fdb0fbeb8894d0b0"],"state_sha256":"8554ac9594f115c7b60cd0535f92168af9bd4d42fdb970c9a85ae3675ade542e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"r3YIwmLEm2dwHP5cQhdDwLi6DhqgOHdmLNcKe2VMv/W4OmdeT+XrRFTJdRjwMTERkMA4EceT04szF3z0XEqXBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-06T19:17:17.753134Z","bundle_sha256":"a1ddde8b98d0aa2d283300adf5e0c0b6fe8da201a3f60b47168ec3fcf39e2c05"}}