{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:TTCKXVRXVU7UERB7HSSQT3ZXP4","short_pith_number":"pith:TTCKXVRX","canonical_record":{"source":{"id":"2403.18742","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-03-27T16:39:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a4eb328f17d3926d8dcb737a6fbc01e121533dc135eba68c3d7d9c03eaa783ad","abstract_canon_sha256":"8eb2b358081f2366507077af11df9d4c45d72e10ec2a9b7501c4d087ee732cc6"},"schema_version":"1.0"},"canonical_sha256":"9cc4abd637ad3f42443f3ca509ef377f197e3ba5c446c5d0a689fd757d41886b","source":{"kind":"arxiv","id":"2403.18742","version":5},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.18742","created_at":"2026-07-05T08:52:52Z"},{"alias_kind":"arxiv_version","alias_value":"2403.18742v5","created_at":"2026-07-05T08:52:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.18742","created_at":"2026-07-05T08:52:52Z"},{"alias_kind":"pith_short_12","alias_value":"TTCKXVRXVU7U","created_at":"2026-07-05T08:52:52Z"},{"alias_kind":"pith_short_16","alias_value":"TTCKXVRXVU7UERB7","created_at":"2026-07-05T08:52:52Z"},{"alias_kind":"pith_short_8","alias_value":"TTCKXVRX","created_at":"2026-07-05T08:52:52Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:TTCKXVRXVU7UERB7HSSQT3ZXP4","target":"record","payload":{"canonical_record":{"source":{"id":"2403.18742","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-03-27T16:39:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a4eb328f17d3926d8dcb737a6fbc01e121533dc135eba68c3d7d9c03eaa783ad","abstract_canon_sha256":"8eb2b358081f2366507077af11df9d4c45d72e10ec2a9b7501c4d087ee732cc6"},"schema_version":"1.0"},"canonical_sha256":"9cc4abd637ad3f42443f3ca509ef377f197e3ba5c446c5d0a689fd757d41886b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:52:52.924603Z","signature_b64":"Fn9MiW1a7zHn1bP0+d6ZO2DBDpk9mE7Mqmg0vUQw3aU/oQM6gtlQTRCnkPlf2J+23wsrM9S1chUEZ2Ickpd9AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9cc4abd637ad3f42443f3ca509ef377f197e3ba5c446c5d0a689fd757d41886b","last_reissued_at":"2026-07-05T08:52:52.924163Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:52:52.924163Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2403.18742","source_version":5,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:52:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Q4TSKkJ7IyvS5ayoxpoDRefg7mi0RudaV5/MvD+l7xUujTxaqRH7FVic/SkDiXtSPQkWEhN4naH5/yP+E4l7CA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T17:39:38.086451Z"},"content_sha256":"804ca2577ccd356a1a0dc935e8dfd9b61f95f8e2cb4bf9979901bb2ca489b6d6","schema_version":"1.0","event_id":"sha256:804ca2577ccd356a1a0dc935e8dfd9b61f95f8e2cb4bf9979901bb2ca489b6d6"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:TTCKXVRXVU7UERB7HSSQT3ZXP4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Understanding the Learning Dynamics of Alignment with Human Feedback","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Shawn Im, Yixuan Li","submitted_at":"2024-03-27T16:39:28Z","abstract_excerpt":"Aligning large language models (LLMs) with human intentions has become a critical task for safely deploying models in real-world systems. While existing alignment approaches have seen empirical success, theoretically understanding how these methods affect model behavior remains an open question. Our work provides an initial attempt to theoretically analyze the learning dynamics of human preference alignment. We formally show how the distribution of preference datasets influences the rate of model updates and provide rigorous guarantees on the training accuracy. Our theory also reveals an intri"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.18742","kind":"arxiv","version":5},"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/2403.18742/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":"2026-07-05T08:52:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dCsx8BapZoCjg+ndZdNeiRVjb5mYy42AW/8UXoTHR1KIAQok/v+CUzS22VBmwd4jYMkVzM/Vl17eSmQOI5Y4CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T17:39:38.087369Z"},"content_sha256":"40f48d465083f07d49efdf83c16ec56fec74c26500730d88a7ebdca1fca112ee","schema_version":"1.0","event_id":"sha256:40f48d465083f07d49efdf83c16ec56fec74c26500730d88a7ebdca1fca112ee"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/TTCKXVRXVU7UERB7HSSQT3ZXP4/bundle.json","state_url":"https://pith.science/pith/TTCKXVRXVU7UERB7HSSQT3ZXP4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/TTCKXVRXVU7UERB7HSSQT3ZXP4/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":"2026-08-07T17:39:38Z","links":{"resolver":"https://pith.science/pith/TTCKXVRXVU7UERB7HSSQT3ZXP4","bundle":"https://pith.science/pith/TTCKXVRXVU7UERB7HSSQT3ZXP4/bundle.json","state":"https://pith.science/pith/TTCKXVRXVU7UERB7HSSQT3ZXP4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/TTCKXVRXVU7UERB7HSSQT3ZXP4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:TTCKXVRXVU7UERB7HSSQT3ZXP4","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":"8eb2b358081f2366507077af11df9d4c45d72e10ec2a9b7501c4d087ee732cc6","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-03-27T16:39:28Z","title_canon_sha256":"a4eb328f17d3926d8dcb737a6fbc01e121533dc135eba68c3d7d9c03eaa783ad"},"schema_version":"1.0","source":{"id":"2403.18742","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.18742","created_at":"2026-07-05T08:52:52Z"},{"alias_kind":"arxiv_version","alias_value":"2403.18742v5","created_at":"2026-07-05T08:52:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.18742","created_at":"2026-07-05T08:52:52Z"},{"alias_kind":"pith_short_12","alias_value":"TTCKXVRXVU7U","created_at":"2026-07-05T08:52:52Z"},{"alias_kind":"pith_short_16","alias_value":"TTCKXVRXVU7UERB7","created_at":"2026-07-05T08:52:52Z"},{"alias_kind":"pith_short_8","alias_value":"TTCKXVRX","created_at":"2026-07-05T08:52:52Z"}],"graph_snapshots":[{"event_id":"sha256:40f48d465083f07d49efdf83c16ec56fec74c26500730d88a7ebdca1fca112ee","target":"graph","created_at":"2026-07-05T08:52:52Z","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/2403.18742/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 intentions has become a critical task for safely deploying models in real-world systems. While existing alignment approaches have seen empirical success, theoretically understanding how these methods affect model behavior remains an open question. Our work provides an initial attempt to theoretically analyze the learning dynamics of human preference alignment. We formally show how the distribution of preference datasets influences the rate of model updates and provide rigorous guarantees on the training accuracy. Our theory also reveals an intri","authors_text":"Shawn Im, Yixuan Li","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-03-27T16:39:28Z","title":"Understanding the Learning Dynamics of Alignment with Human Feedback"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.18742","kind":"arxiv","version":5},"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":"sha256:804ca2577ccd356a1a0dc935e8dfd9b61f95f8e2cb4bf9979901bb2ca489b6d6","target":"record","created_at":"2026-07-05T08:52:52Z","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":"8eb2b358081f2366507077af11df9d4c45d72e10ec2a9b7501c4d087ee732cc6","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-03-27T16:39:28Z","title_canon_sha256":"a4eb328f17d3926d8dcb737a6fbc01e121533dc135eba68c3d7d9c03eaa783ad"},"schema_version":"1.0","source":{"id":"2403.18742","kind":"arxiv","version":5}},"canonical_sha256":"9cc4abd637ad3f42443f3ca509ef377f197e3ba5c446c5d0a689fd757d41886b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9cc4abd637ad3f42443f3ca509ef377f197e3ba5c446c5d0a689fd757d41886b","first_computed_at":"2026-07-05T08:52:52.924163Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:52:52.924163Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Fn9MiW1a7zHn1bP0+d6ZO2DBDpk9mE7Mqmg0vUQw3aU/oQM6gtlQTRCnkPlf2J+23wsrM9S1chUEZ2Ickpd9AQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:52:52.924603Z","signed_message":"canonical_sha256_bytes"},"source_id":"2403.18742","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:804ca2577ccd356a1a0dc935e8dfd9b61f95f8e2cb4bf9979901bb2ca489b6d6","sha256:40f48d465083f07d49efdf83c16ec56fec74c26500730d88a7ebdca1fca112ee"],"state_sha256":"0df07647907bc6f069edd0bc52b2868687313759e61655a0695cadf235e26b95"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"JcRhBEFt/TLPBjXwBFV1LCCu9Em+6bfz3TFsub4R/e/+vLxK2jJzVDH9Y05wvuuGi/ymONsJPJbp5qXZLA4tDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T17:39:38.093320Z","bundle_sha256":"86e79837ee5f6b93fa913768be37096471ab340a6780fefc92948b898a2ada8b"}}