{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:YZBZWDCWVMHRDB23WWZZQZDRWE","short_pith_number":"pith:YZBZWDCW","schema_version":"1.0","canonical_sha256":"c6439b0c56ab0f11875bb5b3986471b11dee4ecabc7e87816c4c26d05e577109","source":{"kind":"arxiv","id":"2601.07525","version":2},"attestation_state":"computed","paper":{"title":"Thinking Before Constraining: A Unified Decoding Framework for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Alonso Silva, Armen Aghasaryan, Laith Zumot, Liubov Tupikina, Mehwish Alam, Ngoc Trinh Hung Nguyen","submitted_at":"2026-01-12T13:25:28Z","abstract_excerpt":"Natural generation allows Large Language Models (LLMs) to produce free-form responses with rich reasoning, yet the lack of structure makes outputs difficult to verify. Conversely, constrained decoding ensures standardized formats but can inadvertently restrict reasoning capabilities by imposing constraints too early in the generation process. We propose a hybrid approach, namely In-Writing, that combines free-form reasoning and structured generation in a single call. The model first performs unconstrained reasoning and only applies structured decoding after a trigger token is generated, explic"},"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":"2601.07525","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-01-12T13:25:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"f48852eefcdbf6b2985d83060e64fbc84b6a75c8fc039eba35e30f8f86955114","abstract_canon_sha256":"b14f214b30c566468828db9b46e1ce7ed8e04887c35690e863898d4ca2aee86f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T02:05:39.354220Z","signature_b64":"HlrTQiMNLULZefMaK3qng7q/ugEOLo6xzRvY8jz3LNHKpRItCDXTgKcFpE98JMGTOlo/YzvNjKRHC15Tms7HDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c6439b0c56ab0f11875bb5b3986471b11dee4ecabc7e87816c4c26d05e577109","last_reissued_at":"2026-05-29T02:05:39.353593Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T02:05:39.353593Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Thinking Before Constraining: A Unified Decoding Framework for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Alonso Silva, Armen Aghasaryan, Laith Zumot, Liubov Tupikina, Mehwish Alam, Ngoc Trinh Hung Nguyen","submitted_at":"2026-01-12T13:25:28Z","abstract_excerpt":"Natural generation allows Large Language Models (LLMs) to produce free-form responses with rich reasoning, yet the lack of structure makes outputs difficult to verify. Conversely, constrained decoding ensures standardized formats but can inadvertently restrict reasoning capabilities by imposing constraints too early in the generation process. We propose a hybrid approach, namely In-Writing, that combines free-form reasoning and structured generation in a single call. The model first performs unconstrained reasoning and only applies structured decoding after a trigger token is generated, explic"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.07525","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/2601.07525/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":"2601.07525","created_at":"2026-05-29T02:05:39.353675+00:00"},{"alias_kind":"arxiv_version","alias_value":"2601.07525v2","created_at":"2026-05-29T02:05:39.353675+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.07525","created_at":"2026-05-29T02:05:39.353675+00:00"},{"alias_kind":"pith_short_12","alias_value":"YZBZWDCWVMHR","created_at":"2026-05-29T02:05:39.353675+00:00"},{"alias_kind":"pith_short_16","alias_value":"YZBZWDCWVMHRDB23","created_at":"2026-05-29T02:05:39.353675+00:00"},{"alias_kind":"pith_short_8","alias_value":"YZBZWDCW","created_at":"2026-05-29T02:05:39.353675+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2604.14862","citing_title":"Schema Key Wording as an Instruction Channel in Structured Generation under Constrained Decoding","ref_index":24,"is_internal_anchor":true},{"citing_arxiv_id":"2605.02363","citing_title":"When Correct Isn't Usable: Improving Structured Output Reliability in Small Language Models","ref_index":15,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YZBZWDCWVMHRDB23WWZZQZDRWE","json":"https://pith.science/pith/YZBZWDCWVMHRDB23WWZZQZDRWE.json","graph_json":"https://pith.science/api/pith-number/YZBZWDCWVMHRDB23WWZZQZDRWE/graph.json","events_json":"https://pith.science/api/pith-number/YZBZWDCWVMHRDB23WWZZQZDRWE/events.json","paper":"https://pith.science/paper/YZBZWDCW"},"agent_actions":{"view_html":"https://pith.science/pith/YZBZWDCWVMHRDB23WWZZQZDRWE","download_json":"https://pith.science/pith/YZBZWDCWVMHRDB23WWZZQZDRWE.json","view_paper":"https://pith.science/paper/YZBZWDCW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2601.07525&json=true","fetch_graph":"https://pith.science/api/pith-number/YZBZWDCWVMHRDB23WWZZQZDRWE/graph.json","fetch_events":"https://pith.science/api/pith-number/YZBZWDCWVMHRDB23WWZZQZDRWE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YZBZWDCWVMHRDB23WWZZQZDRWE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YZBZWDCWVMHRDB23WWZZQZDRWE/action/storage_attestation","attest_author":"https://pith.science/pith/YZBZWDCWVMHRDB23WWZZQZDRWE/action/author_attestation","sign_citation":"https://pith.science/pith/YZBZWDCWVMHRDB23WWZZQZDRWE/action/citation_signature","submit_replication":"https://pith.science/pith/YZBZWDCWVMHRDB23WWZZQZDRWE/action/replication_record"}},"created_at":"2026-05-29T02:05:39.353675+00:00","updated_at":"2026-05-29T02:05:39.353675+00:00"}