{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:SPMAZL3ZDOT2BMMARGHRUTYXZB","short_pith_number":"pith:SPMAZL3Z","canonical_record":{"source":{"id":"2307.12950","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-07-24T17:23:22Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"97b0423404295a67c7539607a88be710353b59282e6957bd2800462162e91b63","abstract_canon_sha256":"6b7bc169393e97ffea1ca91f33e463a8dc537eb09ed3e51bd8c199958e13f6f3"},"schema_version":"1.0"},"canonical_sha256":"93d80caf791ba7a0b180898f1a4f17c87b0e875deefc2225caaf0306ff39cafb","source":{"kind":"arxiv","id":"2307.12950","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2307.12950","created_at":"2026-07-05T07:56:50Z"},{"alias_kind":"arxiv_version","alias_value":"2307.12950v3","created_at":"2026-07-05T07:56:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2307.12950","created_at":"2026-07-05T07:56:50Z"},{"alias_kind":"pith_short_12","alias_value":"SPMAZL3ZDOT2","created_at":"2026-07-05T07:56:50Z"},{"alias_kind":"pith_short_16","alias_value":"SPMAZL3ZDOT2BMMA","created_at":"2026-07-05T07:56:50Z"},{"alias_kind":"pith_short_8","alias_value":"SPMAZL3Z","created_at":"2026-07-05T07:56:50Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:SPMAZL3ZDOT2BMMARGHRUTYXZB","target":"record","payload":{"canonical_record":{"source":{"id":"2307.12950","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-07-24T17:23:22Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"97b0423404295a67c7539607a88be710353b59282e6957bd2800462162e91b63","abstract_canon_sha256":"6b7bc169393e97ffea1ca91f33e463a8dc537eb09ed3e51bd8c199958e13f6f3"},"schema_version":"1.0"},"canonical_sha256":"93d80caf791ba7a0b180898f1a4f17c87b0e875deefc2225caaf0306ff39cafb","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:56:50.732935Z","signature_b64":"hdl32qBzubNwoplJKRv5yGkfLvA9iGcyLx9DKTHIyO4C1nM+ZHSPSmGD1TgBcFzSn1fMdXTQKYzkvIfb4K1oDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"93d80caf791ba7a0b180898f1a4f17c87b0e875deefc2225caaf0306ff39cafb","last_reissued_at":"2026-07-05T07:56:50.732460Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:56:50.732460Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2307.12950","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T07:56:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"t1h3YIv2bSL+cMyxWH0N1ayUd0lbZKL1zJGbhjSqcllNjf9znWUjEa85pXdeO7hCYDGqiIvPhDcYMoqWrBC7Bg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T13:02:53.378443Z"},"content_sha256":"e63b599d1be792273d4c261228fc227df1b6f3f145bda98dd5ba5185d5699434","schema_version":"1.0","event_id":"sha256:e63b599d1be792273d4c261228fc227df1b6f3f145bda98dd5ba5185d5699434"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:SPMAZL3ZDOT2BMMARGHRUTYXZB","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"RLCD: Reinforcement Learning from Contrastive Distillation for Language Model Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Asli Celikyilmaz, Dan Klein, Kevin Yang, Nanyun Peng, Yuandong Tian","submitted_at":"2023-07-24T17:23:22Z","abstract_excerpt":"We propose Reinforcement Learning from Contrastive Distillation (RLCD), a method for aligning language models to follow principles expressed in natural language (e.g., to be more harmless) without using human feedback. RLCD creates preference pairs from two contrasting model outputs, one using a positive prompt designed to encourage following the given principles, and one using a negative prompt designed to encourage violating them. Using two different prompts causes model outputs to be more differentiated on average, resulting in cleaner preference labels in the absence of human annotations. "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2307.12950","kind":"arxiv","version":3},"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/2307.12950/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":"2026-07-05T07:56:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"yIuj/Z0TqZeD9p9vwY9kxaBdqBB7vE5mGR7LVldakuX/jNRHplSpZ+PgUSQ7tA0Q6keSrnB5Rt2fwCm03kzzAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T13:02:53.379039Z"},"content_sha256":"f30edad50d282f5011482b98e7d96c783e0ec55149b8a9f385f05c25f72f87d8","schema_version":"1.0","event_id":"sha256:f30edad50d282f5011482b98e7d96c783e0ec55149b8a9f385f05c25f72f87d8"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/SPMAZL3ZDOT2BMMARGHRUTYXZB/bundle.json","state_url":"https://pith.science/pith/SPMAZL3ZDOT2BMMARGHRUTYXZB/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/SPMAZL3ZDOT2BMMARGHRUTYXZB/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":"2026-08-06T13:02:53Z","links":{"resolver":"https://pith.science/pith/SPMAZL3ZDOT2BMMARGHRUTYXZB","bundle":"https://pith.science/pith/SPMAZL3ZDOT2BMMARGHRUTYXZB/bundle.json","state":"https://pith.science/pith/SPMAZL3ZDOT2BMMARGHRUTYXZB/state.json","well_known_bundle":"https://pith.science/.well-known/pith/SPMAZL3ZDOT2BMMARGHRUTYXZB/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:SPMAZL3ZDOT2BMMARGHRUTYXZB","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":"6b7bc169393e97ffea1ca91f33e463a8dc537eb09ed3e51bd8c199958e13f6f3","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-07-24T17:23:22Z","title_canon_sha256":"97b0423404295a67c7539607a88be710353b59282e6957bd2800462162e91b63"},"schema_version":"1.0","source":{"id":"2307.12950","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2307.12950","created_at":"2026-07-05T07:56:50Z"},{"alias_kind":"arxiv_version","alias_value":"2307.12950v3","created_at":"2026-07-05T07:56:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2307.12950","created_at":"2026-07-05T07:56:50Z"},{"alias_kind":"pith_short_12","alias_value":"SPMAZL3ZDOT2","created_at":"2026-07-05T07:56:50Z"},{"alias_kind":"pith_short_16","alias_value":"SPMAZL3ZDOT2BMMA","created_at":"2026-07-05T07:56:50Z"},{"alias_kind":"pith_short_8","alias_value":"SPMAZL3Z","created_at":"2026-07-05T07:56:50Z"}],"graph_snapshots":[{"event_id":"sha256:f30edad50d282f5011482b98e7d96c783e0ec55149b8a9f385f05c25f72f87d8","target":"graph","created_at":"2026-07-05T07:56:50Z","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/2307.12950/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We propose Reinforcement Learning from Contrastive Distillation (RLCD), a method for aligning language models to follow principles expressed in natural language (e.g., to be more harmless) without using human feedback. RLCD creates preference pairs from two contrasting model outputs, one using a positive prompt designed to encourage following the given principles, and one using a negative prompt designed to encourage violating them. Using two different prompts causes model outputs to be more differentiated on average, resulting in cleaner preference labels in the absence of human annotations. ","authors_text":"Asli Celikyilmaz, Dan Klein, Kevin Yang, Nanyun Peng, Yuandong Tian","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-07-24T17:23:22Z","title":"RLCD: Reinforcement Learning from Contrastive Distillation for Language Model Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2307.12950","kind":"arxiv","version":3},"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:e63b599d1be792273d4c261228fc227df1b6f3f145bda98dd5ba5185d5699434","target":"record","created_at":"2026-07-05T07:56:50Z","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":"6b7bc169393e97ffea1ca91f33e463a8dc537eb09ed3e51bd8c199958e13f6f3","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-07-24T17:23:22Z","title_canon_sha256":"97b0423404295a67c7539607a88be710353b59282e6957bd2800462162e91b63"},"schema_version":"1.0","source":{"id":"2307.12950","kind":"arxiv","version":3}},"canonical_sha256":"93d80caf791ba7a0b180898f1a4f17c87b0e875deefc2225caaf0306ff39cafb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"93d80caf791ba7a0b180898f1a4f17c87b0e875deefc2225caaf0306ff39cafb","first_computed_at":"2026-07-05T07:56:50.732460Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:56:50.732460Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"hdl32qBzubNwoplJKRv5yGkfLvA9iGcyLx9DKTHIyO4C1nM+ZHSPSmGD1TgBcFzSn1fMdXTQKYzkvIfb4K1oDA==","signature_status":"signed_v1","signed_at":"2026-07-05T07:56:50.732935Z","signed_message":"canonical_sha256_bytes"},"source_id":"2307.12950","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e63b599d1be792273d4c261228fc227df1b6f3f145bda98dd5ba5185d5699434","sha256:f30edad50d282f5011482b98e7d96c783e0ec55149b8a9f385f05c25f72f87d8"],"state_sha256":"5019e484fa589d552c2a5f1f73c2bc59ab27b661f91755d6899b9f9efac6590c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7X2efmM8RqUW8l4PXusYu1ebAMF61s8PbnIRH8RKMPns6hQADdlJHS5apgSYqjITAhYrOrK1ABY+y+fUnza+Dw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T13:02:53.382498Z","bundle_sha256":"46318f7a63aad03b9574cab64d6da8fc38308ee050d2b6600a4044383021aa7e"}}