{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:H5IJHVHMJRWSFPMOX3ROB3WKTL","short_pith_number":"pith:H5IJHVHM","schema_version":"1.0","canonical_sha256":"3f5093d4ec4c6d22bd8ebee2e0eeca9af8754b0534325b00da1cc738eff409f9","source":{"kind":"arxiv","id":"2607.28166","version":1},"attestation_state":"computed","paper":{"title":"Where and When to Commit: Candidate-Aware Decoding for Diffusion Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chia-Ming Lee, Chih-Chung Hsu, Ming-Ching Chang, Xin Li, Yu-Lun Liu","submitted_at":"2026-07-30T13:04:47Z","abstract_excerpt":"Diffusion language models (DLMs) expose a provisional prediction at every denoising step, creating an opportunity for generation-time early exit that stops decoding before the schedule is exhausted. Existing early-exit gates decide termination from fixed-region confidence statistics or schedule-dependent rules, evidence too coarse for a decision that freezes every remaining position at once, so they fire prematurely on long chain-of-thought outputs whose answers stabilize only near the end. Adaptive sampling, the other axis of training-free acceleration, paces how quickly positions commit whil"},"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":"2607.28166","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-30T13:04:47Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0924d7061bed2f5003537017d45db95be8f96df2c2605581bf2e5c5bfa66a4cd","abstract_canon_sha256":"c25e847067c1f0c51163a1a8bb23bda5fc810e441798b8ade26ef7f72caf98f3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3f5093d4ec4c6d22bd8ebee2e0eeca9af8754b0534325b00da1cc738eff409f9","last_reissued_at":"2026-07-31T01:36:09.659285Z","signature_status":"unsigned_v0","first_computed_at":"2026-07-31T01:36:09.659285Z"},"graph_snapshot":{"paper":{"title":"Where and When to Commit: Candidate-Aware Decoding for Diffusion Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chia-Ming Lee, Chih-Chung Hsu, Ming-Ching Chang, Xin Li, Yu-Lun Liu","submitted_at":"2026-07-30T13:04:47Z","abstract_excerpt":"Diffusion language models (DLMs) expose a provisional prediction at every denoising step, creating an opportunity for generation-time early exit that stops decoding before the schedule is exhausted. Existing early-exit gates decide termination from fixed-region confidence statistics or schedule-dependent rules, evidence too coarse for a decision that freezes every remaining position at once, so they fire prematurely on long chain-of-thought outputs whose answers stabilize only near the end. Adaptive sampling, the other axis of training-free acceleration, paces how quickly positions commit whil"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.28166","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/2607.28166/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":"2607.28166","created_at":"2026-07-31T01:36:09.662481+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.28166v1","created_at":"2026-07-31T01:36:09.662481+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.28166","created_at":"2026-07-31T01:36:09.662481+00:00"},{"alias_kind":"pith_short_12","alias_value":"H5IJHVHMJRWS","created_at":"2026-07-31T01:36:09.662481+00:00"},{"alias_kind":"pith_short_16","alias_value":"H5IJHVHMJRWSFPMO","created_at":"2026-07-31T01:36:09.662481+00:00"},{"alias_kind":"pith_short_8","alias_value":"H5IJHVHM","created_at":"2026-07-31T01:36:09.662481+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/H5IJHVHMJRWSFPMOX3ROB3WKTL","json":"https://pith.science/pith/H5IJHVHMJRWSFPMOX3ROB3WKTL.json","graph_json":"https://pith.science/api/pith-number/H5IJHVHMJRWSFPMOX3ROB3WKTL/graph.json","events_json":"https://pith.science/api/pith-number/H5IJHVHMJRWSFPMOX3ROB3WKTL/events.json","paper":"https://pith.science/paper/H5IJHVHM"},"agent_actions":{"view_html":"https://pith.science/pith/H5IJHVHMJRWSFPMOX3ROB3WKTL","download_json":"https://pith.science/pith/H5IJHVHMJRWSFPMOX3ROB3WKTL.json","view_paper":"https://pith.science/paper/H5IJHVHM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.28166&json=true","fetch_graph":"https://pith.science/api/pith-number/H5IJHVHMJRWSFPMOX3ROB3WKTL/graph.json","fetch_events":"https://pith.science/api/pith-number/H5IJHVHMJRWSFPMOX3ROB3WKTL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/H5IJHVHMJRWSFPMOX3ROB3WKTL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/H5IJHVHMJRWSFPMOX3ROB3WKTL/action/storage_attestation","attest_author":"https://pith.science/pith/H5IJHVHMJRWSFPMOX3ROB3WKTL/action/author_attestation","sign_citation":"https://pith.science/pith/H5IJHVHMJRWSFPMOX3ROB3WKTL/action/citation_signature","submit_replication":"https://pith.science/pith/H5IJHVHMJRWSFPMOX3ROB3WKTL/action/replication_record"}},"created_at":"2026-07-31T01:36:09.662481+00:00","updated_at":"2026-07-31T01:36:09.662481+00:00"}