{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:WB7EZGELNIKZPILA7I4SSP3YJF","short_pith_number":"pith:WB7EZGEL","schema_version":"1.0","canonical_sha256":"b07e4c988b6a1597a160fa39293f78496487744c383c155ed427fb4600b58021","source":{"kind":"arxiv","id":"2406.04412","version":2},"attestation_state":"computed","paper":{"title":"Spread Preference Annotation: Direct Preference Judgment for Efficient LLM Alignment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Dongyoung Kim, Jaehyung Kim, Jinwoo Shin, Kimin Lee","submitted_at":"2024-06-06T18:01:02Z","abstract_excerpt":"Aligning large language models (LLMs) with human preferences becomes a key component to obtaining state-of-the-art performance, but it yields a huge cost to construct a large human-annotated preference dataset. To tackle this problem, we propose a new framework, Spread Preference Annotation with direct preference judgment (SPA), that boosts the alignment of LLMs using only a very small amount of human-annotated preference data. Our key idea is leveraging the human prior knowledge within the small (seed) data and progressively improving the alignment of LLM, by iteratively generating the respon"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2406.04412","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-06T18:01:02Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"7c10a2d85e2589a81cf648713d07695a17847ae38795887808e9abf56c6e53fb","abstract_canon_sha256":"662e72c08dbe360c771ef645651c1a0723423ab3c8961421d75b3c7cd95c2834"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:23:19.893435Z","signature_b64":"fQ6SL+BzAp6jln3aIxklzzetJMJQmP35ryEUhMASfrvcfmoiaACAC+ZLcm6ElwKoeVY7oWOxXITjHOP3bYZmBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b07e4c988b6a1597a160fa39293f78496487744c383c155ed427fb4600b58021","last_reissued_at":"2026-07-05T10:23:19.892937Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:23:19.892937Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Spread Preference Annotation: Direct Preference Judgment for Efficient LLM Alignment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Dongyoung Kim, Jaehyung Kim, Jinwoo Shin, Kimin Lee","submitted_at":"2024-06-06T18:01:02Z","abstract_excerpt":"Aligning large language models (LLMs) with human preferences becomes a key component to obtaining state-of-the-art performance, but it yields a huge cost to construct a large human-annotated preference dataset. To tackle this problem, we propose a new framework, Spread Preference Annotation with direct preference judgment (SPA), that boosts the alignment of LLMs using only a very small amount of human-annotated preference data. Our key idea is leveraging the human prior knowledge within the small (seed) data and progressively improving the alignment of LLM, by iteratively generating the respon"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.04412","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/2406.04412/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2406.04412","created_at":"2026-07-05T10:23:19.892994+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.04412v2","created_at":"2026-07-05T10:23:19.892994+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.04412","created_at":"2026-07-05T10:23:19.892994+00:00"},{"alias_kind":"pith_short_12","alias_value":"WB7EZGELNIKZ","created_at":"2026-07-05T10:23:19.892994+00:00"},{"alias_kind":"pith_short_16","alias_value":"WB7EZGELNIKZPILA","created_at":"2026-07-05T10:23:19.892994+00:00"},{"alias_kind":"pith_short_8","alias_value":"WB7EZGEL","created_at":"2026-07-05T10:23:19.892994+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.23398","citing_title":"TPMM-DPO: Trajectory-aware Preference-guided Model Merging for Iterative Direct Preference Optimization","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2412.05579","citing_title":"LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methods","ref_index":112,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24536","citing_title":"Generating Place-Based Compromises Between Two Points of View","ref_index":39,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WB7EZGELNIKZPILA7I4SSP3YJF","json":"https://pith.science/pith/WB7EZGELNIKZPILA7I4SSP3YJF.json","graph_json":"https://pith.science/api/pith-number/WB7EZGELNIKZPILA7I4SSP3YJF/graph.json","events_json":"https://pith.science/api/pith-number/WB7EZGELNIKZPILA7I4SSP3YJF/events.json","paper":"https://pith.science/paper/WB7EZGEL"},"agent_actions":{"view_html":"https://pith.science/pith/WB7EZGELNIKZPILA7I4SSP3YJF","download_json":"https://pith.science/pith/WB7EZGELNIKZPILA7I4SSP3YJF.json","view_paper":"https://pith.science/paper/WB7EZGEL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.04412&json=true","fetch_graph":"https://pith.science/api/pith-number/WB7EZGELNIKZPILA7I4SSP3YJF/graph.json","fetch_events":"https://pith.science/api/pith-number/WB7EZGELNIKZPILA7I4SSP3YJF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WB7EZGELNIKZPILA7I4SSP3YJF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WB7EZGELNIKZPILA7I4SSP3YJF/action/storage_attestation","attest_author":"https://pith.science/pith/WB7EZGELNIKZPILA7I4SSP3YJF/action/author_attestation","sign_citation":"https://pith.science/pith/WB7EZGELNIKZPILA7I4SSP3YJF/action/citation_signature","submit_replication":"https://pith.science/pith/WB7EZGELNIKZPILA7I4SSP3YJF/action/replication_record"}},"created_at":"2026-07-05T10:23:19.892994+00:00","updated_at":"2026-07-05T10:23:19.892994+00:00"}