{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:JN5VWVZAUVTVEKYPHNHUASXTU7","short_pith_number":"pith:JN5VWVZA","schema_version":"1.0","canonical_sha256":"4b7b5b5720a567522b0f3b4f404af3a7d47cc3401a0ccf9d34f872cdd26820cb","source":{"kind":"arxiv","id":"2606.02091","version":1},"attestation_state":"computed","paper":{"title":"DFlare: Scaling Up Draft Capacity for Block Diffusion Speculative Decoding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dawei Zhu, Eugene J.Yu, Guanghua Yu, Jianchen Zhu, Jiangshan Duo, Jiebin Zhang, Song Liu, Sujian Li, Weimin Xiong, Yifan Song, Zhenghan Yu, Zheng Li","submitted_at":"2026-06-01T11:18:30Z","abstract_excerpt":"Block diffusion speculative decoding accelerates LLM inference by predicting all tokens within a block simultaneously for the target model to verify in parallel. Predicting an entire block at once requires a sufficiently capable draft model and effective utilization of the target model's internal knowledge. However, the state-of-the-art method DFlash constrains all draft layers to share a single fused representation derived from only a few target layers, limiting per-layer expressiveness and hindering further scaling of draft capacity. In this paper, we present \\modelname, which flares out 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":"2606.02091","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-01T11:18:30Z","cross_cats_sorted":[],"title_canon_sha256":"9829dece1ad0782814687f519a493d1b9c855c99df55d8ff430516b6fc999449","abstract_canon_sha256":"82816d1bd68a052a7bcbff388307b012fd659c1d599bbae854bc4cc195cf457b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:05:06.322108Z","signature_b64":"yp8xtJ1SAzd08zUfqt7f4r8WxCy15myUHFnogY4fxKFCLBS8vSiOoz05QVlswRPJ6CT6SIyDpVmkc7dKtwzyCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4b7b5b5720a567522b0f3b4f404af3a7d47cc3401a0ccf9d34f872cdd26820cb","last_reissued_at":"2026-06-02T02:05:06.321702Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:05:06.321702Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DFlare: Scaling Up Draft Capacity for Block Diffusion Speculative Decoding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dawei Zhu, Eugene J.Yu, Guanghua Yu, Jianchen Zhu, Jiangshan Duo, Jiebin Zhang, Song Liu, Sujian Li, Weimin Xiong, Yifan Song, Zhenghan Yu, Zheng Li","submitted_at":"2026-06-01T11:18:30Z","abstract_excerpt":"Block diffusion speculative decoding accelerates LLM inference by predicting all tokens within a block simultaneously for the target model to verify in parallel. Predicting an entire block at once requires a sufficiently capable draft model and effective utilization of the target model's internal knowledge. However, the state-of-the-art method DFlash constrains all draft layers to share a single fused representation derived from only a few target layers, limiting per-layer expressiveness and hindering further scaling of draft capacity. In this paper, we present \\modelname, which flares out the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.02091","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/2606.02091/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":"2606.02091","created_at":"2026-06-02T02:05:06.321770+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.02091v1","created_at":"2026-06-02T02:05:06.321770+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.02091","created_at":"2026-06-02T02:05:06.321770+00:00"},{"alias_kind":"pith_short_12","alias_value":"JN5VWVZAUVTV","created_at":"2026-06-02T02:05:06.321770+00:00"},{"alias_kind":"pith_short_16","alias_value":"JN5VWVZAUVTVEKYP","created_at":"2026-06-02T02:05:06.321770+00:00"},{"alias_kind":"pith_short_8","alias_value":"JN5VWVZA","created_at":"2026-06-02T02:05:06.321770+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/JN5VWVZAUVTVEKYPHNHUASXTU7","json":"https://pith.science/pith/JN5VWVZAUVTVEKYPHNHUASXTU7.json","graph_json":"https://pith.science/api/pith-number/JN5VWVZAUVTVEKYPHNHUASXTU7/graph.json","events_json":"https://pith.science/api/pith-number/JN5VWVZAUVTVEKYPHNHUASXTU7/events.json","paper":"https://pith.science/paper/JN5VWVZA"},"agent_actions":{"view_html":"https://pith.science/pith/JN5VWVZAUVTVEKYPHNHUASXTU7","download_json":"https://pith.science/pith/JN5VWVZAUVTVEKYPHNHUASXTU7.json","view_paper":"https://pith.science/paper/JN5VWVZA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.02091&json=true","fetch_graph":"https://pith.science/api/pith-number/JN5VWVZAUVTVEKYPHNHUASXTU7/graph.json","fetch_events":"https://pith.science/api/pith-number/JN5VWVZAUVTVEKYPHNHUASXTU7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JN5VWVZAUVTVEKYPHNHUASXTU7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JN5VWVZAUVTVEKYPHNHUASXTU7/action/storage_attestation","attest_author":"https://pith.science/pith/JN5VWVZAUVTVEKYPHNHUASXTU7/action/author_attestation","sign_citation":"https://pith.science/pith/JN5VWVZAUVTVEKYPHNHUASXTU7/action/citation_signature","submit_replication":"https://pith.science/pith/JN5VWVZAUVTVEKYPHNHUASXTU7/action/replication_record"}},"created_at":"2026-06-02T02:05:06.321770+00:00","updated_at":"2026-06-02T02:05:06.321770+00:00"}