{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:SSVUFH5WGY2CKNNYTLTZSO7IN2","short_pith_number":"pith:SSVUFH5W","schema_version":"1.0","canonical_sha256":"94ab429fb636342535b89ae7993be86e83856a59bc7651e092b2822b53930ad8","source":{"kind":"arxiv","id":"2406.13892","version":2},"attestation_state":"computed","paper":{"title":"Adaptable Logical Control for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Guy Van den Broeck, Honghua Zhang, Masahiro Yoshida, Nanyun Peng, Po-Nien Kung","submitted_at":"2024-06-19T23:47:59Z","abstract_excerpt":"Despite the success of Large Language Models (LLMs) on various tasks following human instructions, controlling model generation at inference time poses a persistent challenge. In this paper, we introduce Ctrl-G, an adaptable framework that facilitates tractable and flexible control of LLM generation to reliably follow logical constraints. Ctrl-G combines any production-ready LLM with a Hidden Markov Model, enabling LLM outputs to adhere to logical constraints represented as deterministic finite automata. We show that Ctrl-G, when applied to a TULU2-7B model, outperforms GPT3.5 and GPT4 on the "},"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":"2406.13892","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-19T23:47:59Z","cross_cats_sorted":[],"title_canon_sha256":"597537c5655d0d4a1ac8cc4cae1ce22e6acd5ebfb0b0c17ed3cc48e99a554ff4","abstract_canon_sha256":"702ac1b984c7bd3778467b0360e095cfc26237f32f78a02e03dc8a9eff6c2d05"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:56:21.737503Z","signature_b64":"iJhHpUp4OqrFXLMBd/e+niu7Qe1TSDakTOLTIUMY6sysC8xkmHotw34FKHWaCtsqtCb5KtIZ546DEY8ROAZcAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"94ab429fb636342535b89ae7993be86e83856a59bc7651e092b2822b53930ad8","last_reissued_at":"2026-07-05T08:56:21.736951Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:56:21.736951Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adaptable Logical Control for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Guy Van den Broeck, Honghua Zhang, Masahiro Yoshida, Nanyun Peng, Po-Nien Kung","submitted_at":"2024-06-19T23:47:59Z","abstract_excerpt":"Despite the success of Large Language Models (LLMs) on various tasks following human instructions, controlling model generation at inference time poses a persistent challenge. In this paper, we introduce Ctrl-G, an adaptable framework that facilitates tractable and flexible control of LLM generation to reliably follow logical constraints. Ctrl-G combines any production-ready LLM with a Hidden Markov Model, enabling LLM outputs to adhere to logical constraints represented as deterministic finite automata. We show that Ctrl-G, when applied to a TULU2-7B model, outperforms GPT3.5 and GPT4 on the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.13892","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/2406.13892/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":"2406.13892","created_at":"2026-07-05T08:56:21.737008+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.13892v2","created_at":"2026-07-05T08:56:21.737008+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.13892","created_at":"2026-07-05T08:56:21.737008+00:00"},{"alias_kind":"pith_short_12","alias_value":"SSVUFH5WGY2C","created_at":"2026-07-05T08:56:21.737008+00:00"},{"alias_kind":"pith_short_16","alias_value":"SSVUFH5WGY2CKNNY","created_at":"2026-07-05T08:56:21.737008+00:00"},{"alias_kind":"pith_short_8","alias_value":"SSVUFH5W","created_at":"2026-07-05T08:56:21.737008+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/SSVUFH5WGY2CKNNYTLTZSO7IN2","json":"https://pith.science/pith/SSVUFH5WGY2CKNNYTLTZSO7IN2.json","graph_json":"https://pith.science/api/pith-number/SSVUFH5WGY2CKNNYTLTZSO7IN2/graph.json","events_json":"https://pith.science/api/pith-number/SSVUFH5WGY2CKNNYTLTZSO7IN2/events.json","paper":"https://pith.science/paper/SSVUFH5W"},"agent_actions":{"view_html":"https://pith.science/pith/SSVUFH5WGY2CKNNYTLTZSO7IN2","download_json":"https://pith.science/pith/SSVUFH5WGY2CKNNYTLTZSO7IN2.json","view_paper":"https://pith.science/paper/SSVUFH5W","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.13892&json=true","fetch_graph":"https://pith.science/api/pith-number/SSVUFH5WGY2CKNNYTLTZSO7IN2/graph.json","fetch_events":"https://pith.science/api/pith-number/SSVUFH5WGY2CKNNYTLTZSO7IN2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SSVUFH5WGY2CKNNYTLTZSO7IN2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SSVUFH5WGY2CKNNYTLTZSO7IN2/action/storage_attestation","attest_author":"https://pith.science/pith/SSVUFH5WGY2CKNNYTLTZSO7IN2/action/author_attestation","sign_citation":"https://pith.science/pith/SSVUFH5WGY2CKNNYTLTZSO7IN2/action/citation_signature","submit_replication":"https://pith.science/pith/SSVUFH5WGY2CKNNYTLTZSO7IN2/action/replication_record"}},"created_at":"2026-07-05T08:56:21.737008+00:00","updated_at":"2026-07-05T08:56:21.737008+00:00"}