{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MARYGN2XEXCBQBXO4L7W43Q4NV","short_pith_number":"pith:MARYGN2X","schema_version":"1.0","canonical_sha256":"602383375725c41806eee2ff6e6e1c6d52ac952caa6ddc72953bd062d544fcf0","source":{"kind":"arxiv","id":"2506.07434","version":1},"attestation_state":"computed","paper":{"title":"Well Begun is Half Done: Low-resource Preference Alignment by Weak-to-Strong Decoding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Feifan Song, Guoyin Wang, Houfeng Wang, Shaohang Wei, Tianyu Liu, Wen Luo, Yuxuan Fan","submitted_at":"2025-06-09T05:21:22Z","abstract_excerpt":"Large Language Models (LLMs) require alignment with human preferences to avoid generating offensive, false, or meaningless content. Recently, low-resource methods for LLM alignment have been popular, while still facing challenges in obtaining both high-quality and aligned content. Motivated by the observation that the difficulty of generating aligned responses is concentrated at the beginning of decoding, we propose a novel framework, Weak-to-Strong Decoding (WSD), to enhance the alignment ability of base models by the guidance of a small aligned model. The small model first drafts well-aligne"},"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":"2506.07434","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-06-09T05:21:22Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"1e5a6673123e4a1ca38ec7d349b24f4be0d6a287ca3b9d0ff111d3e297929bd7","abstract_canon_sha256":"bec69696fb64e8d8970d6a0700b12da8af02f31f2e88d61ad71c23d80b1da678"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:18:27.215911Z","signature_b64":"Vax/Ip+d6kEr6eOHvsvWeduo77R7OJyWHXq9EYJoE+1Xg+KHVxiWRLaCCtxz/N8yneuTclRr8RS0U5v/xHLZCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"602383375725c41806eee2ff6e6e1c6d52ac952caa6ddc72953bd062d544fcf0","last_reissued_at":"2026-07-05T11:18:27.215376Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:18:27.215376Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Well Begun is Half Done: Low-resource Preference Alignment by Weak-to-Strong Decoding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Feifan Song, Guoyin Wang, Houfeng Wang, Shaohang Wei, Tianyu Liu, Wen Luo, Yuxuan Fan","submitted_at":"2025-06-09T05:21:22Z","abstract_excerpt":"Large Language Models (LLMs) require alignment with human preferences to avoid generating offensive, false, or meaningless content. Recently, low-resource methods for LLM alignment have been popular, while still facing challenges in obtaining both high-quality and aligned content. Motivated by the observation that the difficulty of generating aligned responses is concentrated at the beginning of decoding, we propose a novel framework, Weak-to-Strong Decoding (WSD), to enhance the alignment ability of base models by the guidance of a small aligned model. The small model first drafts well-aligne"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.07434","kind":"arxiv","version":1},"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/2506.07434/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":"2506.07434","created_at":"2026-07-05T11:18:27.215441+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.07434v1","created_at":"2026-07-05T11:18:27.215441+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.07434","created_at":"2026-07-05T11:18:27.215441+00:00"},{"alias_kind":"pith_short_12","alias_value":"MARYGN2XEXCB","created_at":"2026-07-05T11:18:27.215441+00:00"},{"alias_kind":"pith_short_16","alias_value":"MARYGN2XEXCBQBXO","created_at":"2026-07-05T11:18:27.215441+00:00"},{"alias_kind":"pith_short_8","alias_value":"MARYGN2X","created_at":"2026-07-05T11:18:27.215441+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MARYGN2XEXCBQBXO4L7W43Q4NV","json":"https://pith.science/pith/MARYGN2XEXCBQBXO4L7W43Q4NV.json","graph_json":"https://pith.science/api/pith-number/MARYGN2XEXCBQBXO4L7W43Q4NV/graph.json","events_json":"https://pith.science/api/pith-number/MARYGN2XEXCBQBXO4L7W43Q4NV/events.json","paper":"https://pith.science/paper/MARYGN2X"},"agent_actions":{"view_html":"https://pith.science/pith/MARYGN2XEXCBQBXO4L7W43Q4NV","download_json":"https://pith.science/pith/MARYGN2XEXCBQBXO4L7W43Q4NV.json","view_paper":"https://pith.science/paper/MARYGN2X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.07434&json=true","fetch_graph":"https://pith.science/api/pith-number/MARYGN2XEXCBQBXO4L7W43Q4NV/graph.json","fetch_events":"https://pith.science/api/pith-number/MARYGN2XEXCBQBXO4L7W43Q4NV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MARYGN2XEXCBQBXO4L7W43Q4NV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MARYGN2XEXCBQBXO4L7W43Q4NV/action/storage_attestation","attest_author":"https://pith.science/pith/MARYGN2XEXCBQBXO4L7W43Q4NV/action/author_attestation","sign_citation":"https://pith.science/pith/MARYGN2XEXCBQBXO4L7W43Q4NV/action/citation_signature","submit_replication":"https://pith.science/pith/MARYGN2XEXCBQBXO4L7W43Q4NV/action/replication_record"}},"created_at":"2026-07-05T11:18:27.215441+00:00","updated_at":"2026-07-05T11:18:27.215441+00:00"}