{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:HCDTEGLIXDKLKDKACASB7VSFLX","short_pith_number":"pith:HCDTEGLI","canonical_record":{"source":{"id":"2310.02263","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-03T17:59:46Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"5e79a179bfa452110aa3a953ba76ea9fbbef9b17f41b656930ef8a70e5315028","abstract_canon_sha256":"f7885a6f7f4fcf677721a5f3017a223cc964e1bbf64ebb9485a49f59a7e38e14"},"schema_version":"1.0"},"canonical_sha256":"3887321968b8d4b50d4010241fd6455df0cfc3af1cddd691c40d46f82bd8cab3","source":{"kind":"arxiv","id":"2310.02263","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.02263","created_at":"2026-07-05T08:03:44Z"},{"alias_kind":"arxiv_version","alias_value":"2310.02263v2","created_at":"2026-07-05T08:03:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.02263","created_at":"2026-07-05T08:03:44Z"},{"alias_kind":"pith_short_12","alias_value":"HCDTEGLIXDKL","created_at":"2026-07-05T08:03:44Z"},{"alias_kind":"pith_short_16","alias_value":"HCDTEGLIXDKLKDKA","created_at":"2026-07-05T08:03:44Z"},{"alias_kind":"pith_short_8","alias_value":"HCDTEGLI","created_at":"2026-07-05T08:03:44Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:HCDTEGLIXDKLKDKACASB7VSFLX","target":"record","payload":{"canonical_record":{"source":{"id":"2310.02263","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-03T17:59:46Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"5e79a179bfa452110aa3a953ba76ea9fbbef9b17f41b656930ef8a70e5315028","abstract_canon_sha256":"f7885a6f7f4fcf677721a5f3017a223cc964e1bbf64ebb9485a49f59a7e38e14"},"schema_version":"1.0"},"canonical_sha256":"3887321968b8d4b50d4010241fd6455df0cfc3af1cddd691c40d46f82bd8cab3","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:03:44.584050Z","signature_b64":"KdiBRGlfKrnptmd0kEEziy2tTWrbcCvpSuOU/D6k51qsimKRrzYHJ/OXDAkMi1lzkd/P/zic14oDTIApBdUbDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3887321968b8d4b50d4010241fd6455df0cfc3af1cddd691c40d46f82bd8cab3","last_reissued_at":"2026-07-05T08:03:44.583571Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:03:44.583571Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2310.02263","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":"2026-07-05T08:03:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"yYQfH8I93Ll0GJrhxQJq5JHA+wtCKSNf/KaM2eZN9yzz6QSWrwcZoei2aD53hw++lY+Lj6Iazrv3qQLOgAXxAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-12T14:09:53.243565Z"},"content_sha256":"4d0633d18aab2f0b4b3b7ed38009408b3f0de9b883799ca0890e06ef595aee90","schema_version":"1.0","event_id":"sha256:4d0633d18aab2f0b4b3b7ed38009408b3f0de9b883799ca0890e06ef595aee90"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:HCDTEGLIXDKLKDKACASB7VSFLX","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Automatic Pair Construction for Contrastive Post-training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Ahmed Hassan Awadallah, Canwen Xu, Corby Rosset, Ethan C. Chau, Jennifer Neville, Julian McAuley, Luciano Del Corro, Nikhil Rao, Shweti Mahajan","submitted_at":"2023-10-03T17:59:46Z","abstract_excerpt":"Alignment serves as an important step to steer large language models (LLMs) towards human preferences. In this paper, we propose an automatic way to construct contrastive data for LLM, using preference pairs from multiple models of varying strengths (e.g., InstructGPT, ChatGPT and GPT-4). We compare the contrastive techniques of SLiC and DPO to SFT baselines and find that DPO provides a step-function improvement even after continuing SFT saturates. We also explore a data curriculum learning scheme for contrastive post-training, which starts by learning from \"easier\" pairs and transitioning to "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.02263","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/2310.02263/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":"2026-07-05T08:03:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"m8U5UaXzd1mB+HwndGnGFCV4zxn5qAVAdL5eh351vlxwWET9RQiN+ImA85RWN36FSU/keC9oFsVDjGskb3HoDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-12T14:09:53.243958Z"},"content_sha256":"f556a0c9354f1b69cb5482cfbed58f7c890ae3a003b26697488b2df9918a4ab5","schema_version":"1.0","event_id":"sha256:f556a0c9354f1b69cb5482cfbed58f7c890ae3a003b26697488b2df9918a4ab5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/HCDTEGLIXDKLKDKACASB7VSFLX/bundle.json","state_url":"https://pith.science/pith/HCDTEGLIXDKLKDKACASB7VSFLX/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/HCDTEGLIXDKLKDKACASB7VSFLX/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":"2026-07-12T14:09:53Z","links":{"resolver":"https://pith.science/pith/HCDTEGLIXDKLKDKACASB7VSFLX","bundle":"https://pith.science/pith/HCDTEGLIXDKLKDKACASB7VSFLX/bundle.json","state":"https://pith.science/pith/HCDTEGLIXDKLKDKACASB7VSFLX/state.json","well_known_bundle":"https://pith.science/.well-known/pith/HCDTEGLIXDKLKDKACASB7VSFLX/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:HCDTEGLIXDKLKDKACASB7VSFLX","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":"f7885a6f7f4fcf677721a5f3017a223cc964e1bbf64ebb9485a49f59a7e38e14","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-03T17:59:46Z","title_canon_sha256":"5e79a179bfa452110aa3a953ba76ea9fbbef9b17f41b656930ef8a70e5315028"},"schema_version":"1.0","source":{"id":"2310.02263","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.02263","created_at":"2026-07-05T08:03:44Z"},{"alias_kind":"arxiv_version","alias_value":"2310.02263v2","created_at":"2026-07-05T08:03:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.02263","created_at":"2026-07-05T08:03:44Z"},{"alias_kind":"pith_short_12","alias_value":"HCDTEGLIXDKL","created_at":"2026-07-05T08:03:44Z"},{"alias_kind":"pith_short_16","alias_value":"HCDTEGLIXDKLKDKA","created_at":"2026-07-05T08:03:44Z"},{"alias_kind":"pith_short_8","alias_value":"HCDTEGLI","created_at":"2026-07-05T08:03:44Z"}],"graph_snapshots":[{"event_id":"sha256:f556a0c9354f1b69cb5482cfbed58f7c890ae3a003b26697488b2df9918a4ab5","target":"graph","created_at":"2026-07-05T08:03:44Z","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/2310.02263/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Alignment serves as an important step to steer large language models (LLMs) towards human preferences. In this paper, we propose an automatic way to construct contrastive data for LLM, using preference pairs from multiple models of varying strengths (e.g., InstructGPT, ChatGPT and GPT-4). We compare the contrastive techniques of SLiC and DPO to SFT baselines and find that DPO provides a step-function improvement even after continuing SFT saturates. We also explore a data curriculum learning scheme for contrastive post-training, which starts by learning from \"easier\" pairs and transitioning to ","authors_text":"Ahmed Hassan Awadallah, Canwen Xu, Corby Rosset, Ethan C. Chau, Jennifer Neville, Julian McAuley, Luciano Del Corro, Nikhil Rao, Shweti Mahajan","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-03T17:59:46Z","title":"Automatic Pair Construction for Contrastive Post-training"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.02263","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":"sha256:4d0633d18aab2f0b4b3b7ed38009408b3f0de9b883799ca0890e06ef595aee90","target":"record","created_at":"2026-07-05T08:03:44Z","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":"f7885a6f7f4fcf677721a5f3017a223cc964e1bbf64ebb9485a49f59a7e38e14","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-03T17:59:46Z","title_canon_sha256":"5e79a179bfa452110aa3a953ba76ea9fbbef9b17f41b656930ef8a70e5315028"},"schema_version":"1.0","source":{"id":"2310.02263","kind":"arxiv","version":2}},"canonical_sha256":"3887321968b8d4b50d4010241fd6455df0cfc3af1cddd691c40d46f82bd8cab3","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3887321968b8d4b50d4010241fd6455df0cfc3af1cddd691c40d46f82bd8cab3","first_computed_at":"2026-07-05T08:03:44.583571Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:03:44.583571Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KdiBRGlfKrnptmd0kEEziy2tTWrbcCvpSuOU/D6k51qsimKRrzYHJ/OXDAkMi1lzkd/P/zic14oDTIApBdUbDA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:03:44.584050Z","signed_message":"canonical_sha256_bytes"},"source_id":"2310.02263","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4d0633d18aab2f0b4b3b7ed38009408b3f0de9b883799ca0890e06ef595aee90","sha256:f556a0c9354f1b69cb5482cfbed58f7c890ae3a003b26697488b2df9918a4ab5"],"state_sha256":"71c9b11d2841a9ba3c476b2d84435ee564a9c15da4dd4705801f025167755357"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PmVI9IgiNSNTcUsu4SYsCFExJfYf6FKiWaOO0bE1qP0jjb+LK+WI7OsH0svngBAhh0dWH6ps/0wOhW69jn6dCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-12T14:09:53.246019Z","bundle_sha256":"d9e3cf20ba000a4d23d4788502ac22190e1cff13fc1aa4d28695164b85aae72e"}}