{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BIH56DPC4GYKTYF6DKN2GOBYDI","short_pith_number":"pith:BIH56DPC","schema_version":"1.0","canonical_sha256":"0a0fdf0de2e1b0a9e0be1a9ba338381a36ec6b0f6620e11155717acd94ba9a5b","source":{"kind":"arxiv","id":"2406.17276","version":4},"attestation_state":"computed","paper":{"title":"OPT-Tree: Speculative Decoding with Adaptive Draft Tree Structure","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jikai Wang, Juntao Li, Min Zhang, Qingrong Xia, Xinyu Duan, Yi Su, Zhefeng Wang, Zi Ye","submitted_at":"2024-06-25T04:45:53Z","abstract_excerpt":"Autoregressive language models demonstrate excellent performance in various scenarios. However, the inference efficiency is limited by its one-step-one-word generation mode, which has become a pressing problem recently as the models become increasingly larger. Speculative decoding employs a \"draft and then verify\" mechanism to allow multiple tokens to be generated in one step, realizing lossless acceleration. Existing methods mainly adopt fixed heuristic draft structures, which fail to adapt to different situations to maximize the acceptance length during verification. To alleviate this dilemm"},"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":"2406.17276","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-25T04:45:53Z","cross_cats_sorted":[],"title_canon_sha256":"ff90c0f87ff804fca33f43a4cd8621a87e0f940e6bbd9af1add29a13d88be5a8","abstract_canon_sha256":"144b8b466e3e9e7b997f2caf720b0624e9ef0b3a9e920ec69e76e43c3553156f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:53:10.732518Z","signature_b64":"bkg9inTpEX7pOxgslyGTJ8C5+EWtuzSvwDyfxdGCmAw5qIeH+gpIfNZq+SBr7ySjZDRgK378T7R+LcXqQ/kWDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0a0fdf0de2e1b0a9e0be1a9ba338381a36ec6b0f6620e11155717acd94ba9a5b","last_reissued_at":"2026-07-05T10:53:10.732042Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:53:10.732042Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OPT-Tree: Speculative Decoding with Adaptive Draft Tree Structure","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jikai Wang, Juntao Li, Min Zhang, Qingrong Xia, Xinyu Duan, Yi Su, Zhefeng Wang, Zi Ye","submitted_at":"2024-06-25T04:45:53Z","abstract_excerpt":"Autoregressive language models demonstrate excellent performance in various scenarios. However, the inference efficiency is limited by its one-step-one-word generation mode, which has become a pressing problem recently as the models become increasingly larger. Speculative decoding employs a \"draft and then verify\" mechanism to allow multiple tokens to be generated in one step, realizing lossless acceleration. Existing methods mainly adopt fixed heuristic draft structures, which fail to adapt to different situations to maximize the acceptance length during verification. To alleviate this dilemm"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.17276","kind":"arxiv","version":4},"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/2406.17276/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":"2406.17276","created_at":"2026-07-05T10:53:10.732097+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.17276v4","created_at":"2026-07-05T10:53:10.732097+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.17276","created_at":"2026-07-05T10:53:10.732097+00:00"},{"alias_kind":"pith_short_12","alias_value":"BIH56DPC4GYK","created_at":"2026-07-05T10:53:10.732097+00:00"},{"alias_kind":"pith_short_16","alias_value":"BIH56DPC4GYKTYF6","created_at":"2026-07-05T10:53:10.732097+00:00"},{"alias_kind":"pith_short_8","alias_value":"BIH56DPC","created_at":"2026-07-05T10:53:10.732097+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.12303","citing_title":"From 2D Grids to 1D Tokens: Reforming Shared Representations for Multimodal Image Fusion","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00535","citing_title":"DREAM-S: Speculative Decoding with Searchable Drafting and Target-Aware Refinement for Multimodal Generation","ref_index":99,"is_internal_anchor":false},{"citing_arxiv_id":"2507.01449","citing_title":"LogitSpec: Accelerating Retrieval-based Speculative Decoding via Next Next Token Speculation","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11186","citing_title":"CATS: Cascaded Adaptive Tree Speculation for Memory-Limited LLM Inference Acceleration","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BIH56DPC4GYKTYF6DKN2GOBYDI","json":"https://pith.science/pith/BIH56DPC4GYKTYF6DKN2GOBYDI.json","graph_json":"https://pith.science/api/pith-number/BIH56DPC4GYKTYF6DKN2GOBYDI/graph.json","events_json":"https://pith.science/api/pith-number/BIH56DPC4GYKTYF6DKN2GOBYDI/events.json","paper":"https://pith.science/paper/BIH56DPC"},"agent_actions":{"view_html":"https://pith.science/pith/BIH56DPC4GYKTYF6DKN2GOBYDI","download_json":"https://pith.science/pith/BIH56DPC4GYKTYF6DKN2GOBYDI.json","view_paper":"https://pith.science/paper/BIH56DPC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.17276&json=true","fetch_graph":"https://pith.science/api/pith-number/BIH56DPC4GYKTYF6DKN2GOBYDI/graph.json","fetch_events":"https://pith.science/api/pith-number/BIH56DPC4GYKTYF6DKN2GOBYDI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BIH56DPC4GYKTYF6DKN2GOBYDI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BIH56DPC4GYKTYF6DKN2GOBYDI/action/storage_attestation","attest_author":"https://pith.science/pith/BIH56DPC4GYKTYF6DKN2GOBYDI/action/author_attestation","sign_citation":"https://pith.science/pith/BIH56DPC4GYKTYF6DKN2GOBYDI/action/citation_signature","submit_replication":"https://pith.science/pith/BIH56DPC4GYKTYF6DKN2GOBYDI/action/replication_record"}},"created_at":"2026-07-05T10:53:10.732097+00:00","updated_at":"2026-07-05T10:53:10.732097+00:00"}