{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:WFAW2UUCLS7GRMVHDEAT6NUPFN","short_pith_number":"pith:WFAW2UUC","schema_version":"1.0","canonical_sha256":"b1416d52825cbe68b2a719013f368f2b7b0a73129153f75bbfcf06ef1972b696","source":{"kind":"arxiv","id":"2409.17819","version":1},"attestation_state":"computed","paper":{"title":"Inference-Time Language Model Alignment via Integrated Value Guidance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chao Yang, Yuanfu Wang, Yu Qiao, Zhanhui Zhou, Zhixuan Liu","submitted_at":"2024-09-26T13:15:18Z","abstract_excerpt":"Large language models are typically fine-tuned to align with human preferences, but tuning large models is computationally intensive and complex. In this work, we introduce $\\textit{Integrated Value Guidance}$ (IVG), a method that uses implicit and explicit value functions to guide language model decoding at token and chunk-level respectively, efficiently aligning large language models purely at inference time. This approach circumvents the complexities of direct fine-tuning and outperforms traditional methods. Empirically, we demonstrate the versatility of IVG across various tasks. In control"},"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":"2409.17819","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-26T13:15:18Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"4c7ced2f1350584ec580b79ab9706313566018dc23f5d910beeb26bc24b0be22","abstract_canon_sha256":"4c10dbef47f6df63bc8a5e801c04a6359a9288f5379cce7b7e09d3c14ac8f946"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:12:16.183952Z","signature_b64":"6oWaW795Zsmk1RekzZlD540Lde/+hDqJKYjeVAwPtGivc91tMzmcgKBVG9yGuOkYDjriMkbq3VN2mZL/hf8cCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b1416d52825cbe68b2a719013f368f2b7b0a73129153f75bbfcf06ef1972b696","last_reissued_at":"2026-07-05T09:12:16.183471Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:12:16.183471Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Inference-Time Language Model Alignment via Integrated Value Guidance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chao Yang, Yuanfu Wang, Yu Qiao, Zhanhui Zhou, Zhixuan Liu","submitted_at":"2024-09-26T13:15:18Z","abstract_excerpt":"Large language models are typically fine-tuned to align with human preferences, but tuning large models is computationally intensive and complex. In this work, we introduce $\\textit{Integrated Value Guidance}$ (IVG), a method that uses implicit and explicit value functions to guide language model decoding at token and chunk-level respectively, efficiently aligning large language models purely at inference time. This approach circumvents the complexities of direct fine-tuning and outperforms traditional methods. Empirically, we demonstrate the versatility of IVG across various tasks. In control"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.17819","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/2409.17819/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":"2409.17819","created_at":"2026-07-05T09:12:16.183529+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.17819v1","created_at":"2026-07-05T09:12:16.183529+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.17819","created_at":"2026-07-05T09:12:16.183529+00:00"},{"alias_kind":"pith_short_12","alias_value":"WFAW2UUCLS7G","created_at":"2026-07-05T09:12:16.183529+00:00"},{"alias_kind":"pith_short_16","alias_value":"WFAW2UUCLS7GRMVH","created_at":"2026-07-05T09:12:16.183529+00:00"},{"alias_kind":"pith_short_8","alias_value":"WFAW2UUC","created_at":"2026-07-05T09:12:16.183529+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/WFAW2UUCLS7GRMVHDEAT6NUPFN","json":"https://pith.science/pith/WFAW2UUCLS7GRMVHDEAT6NUPFN.json","graph_json":"https://pith.science/api/pith-number/WFAW2UUCLS7GRMVHDEAT6NUPFN/graph.json","events_json":"https://pith.science/api/pith-number/WFAW2UUCLS7GRMVHDEAT6NUPFN/events.json","paper":"https://pith.science/paper/WFAW2UUC"},"agent_actions":{"view_html":"https://pith.science/pith/WFAW2UUCLS7GRMVHDEAT6NUPFN","download_json":"https://pith.science/pith/WFAW2UUCLS7GRMVHDEAT6NUPFN.json","view_paper":"https://pith.science/paper/WFAW2UUC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.17819&json=true","fetch_graph":"https://pith.science/api/pith-number/WFAW2UUCLS7GRMVHDEAT6NUPFN/graph.json","fetch_events":"https://pith.science/api/pith-number/WFAW2UUCLS7GRMVHDEAT6NUPFN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WFAW2UUCLS7GRMVHDEAT6NUPFN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WFAW2UUCLS7GRMVHDEAT6NUPFN/action/storage_attestation","attest_author":"https://pith.science/pith/WFAW2UUCLS7GRMVHDEAT6NUPFN/action/author_attestation","sign_citation":"https://pith.science/pith/WFAW2UUCLS7GRMVHDEAT6NUPFN/action/citation_signature","submit_replication":"https://pith.science/pith/WFAW2UUCLS7GRMVHDEAT6NUPFN/action/replication_record"}},"created_at":"2026-07-05T09:12:16.183529+00:00","updated_at":"2026-07-05T09:12:16.183529+00:00"}