{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:KV237ZSZJIXU5YUMREBM2AV4OJ","short_pith_number":"pith:KV237ZSZ","schema_version":"1.0","canonical_sha256":"5575bfe6594a2f4ee28c8902cd02bc7259e3a8b7aa8286f136e31b6ea6a253da","source":{"kind":"arxiv","id":"2411.11055","version":3},"attestation_state":"computed","paper":{"title":"FastDraft: How to Train Your Draft","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dorin Shteyman, Guy Boudoukh, Igor Margulis, Ofir Zafrir, Shira Guskin","submitted_at":"2024-11-17T12:32:44Z","abstract_excerpt":"Speculative Decoding has gained popularity as an effective technique for accelerating the auto-regressive inference process of Large Language Models. However, Speculative Decoding entirely relies on the availability of efficient draft models, which are often lacking for many existing language models due to a stringent constraint of vocabulary compatibility. In this work we introduce FastDraft, a novel and efficient approach for pre-training and aligning a draft model to any large language model by incorporating efficient pre-training, followed by fine-tuning over synthetic datasets generated b"},"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":"2411.11055","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-11-17T12:32:44Z","cross_cats_sorted":[],"title_canon_sha256":"f1f70b0adad4c83f1f28206f87afcef70a6553aefce22c05d7217d839a3983ea","abstract_canon_sha256":"26cde31dc20532cc189126c5d125f9508a98b7411a64ef130e2f7887b270a789"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:16:04.662331Z","signature_b64":"54N+RfMD0YTzT9SRn8BPGKDSnSTZeXkR0UYSIM38otxA0EFQF3Cd8N+u6bhEEUZnF2Qaz1aqeecvc7h03TybDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5575bfe6594a2f4ee28c8902cd02bc7259e3a8b7aa8286f136e31b6ea6a253da","last_reissued_at":"2026-07-05T11:16:04.661549Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:16:04.661549Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FastDraft: How to Train Your Draft","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dorin Shteyman, Guy Boudoukh, Igor Margulis, Ofir Zafrir, Shira Guskin","submitted_at":"2024-11-17T12:32:44Z","abstract_excerpt":"Speculative Decoding has gained popularity as an effective technique for accelerating the auto-regressive inference process of Large Language Models. However, Speculative Decoding entirely relies on the availability of efficient draft models, which are often lacking for many existing language models due to a stringent constraint of vocabulary compatibility. In this work we introduce FastDraft, a novel and efficient approach for pre-training and aligning a draft model to any large language model by incorporating efficient pre-training, followed by fine-tuning over synthetic datasets generated b"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.11055","kind":"arxiv","version":3},"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/2411.11055/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":"2411.11055","created_at":"2026-07-05T11:16:04.661613+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.11055v3","created_at":"2026-07-05T11:16:04.661613+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.11055","created_at":"2026-07-05T11:16:04.661613+00:00"},{"alias_kind":"pith_short_12","alias_value":"KV237ZSZJIXU","created_at":"2026-07-05T11:16:04.661613+00:00"},{"alias_kind":"pith_short_16","alias_value":"KV237ZSZJIXU5YUM","created_at":"2026-07-05T11:16:04.661613+00:00"},{"alias_kind":"pith_short_8","alias_value":"KV237ZSZ","created_at":"2026-07-05T11:16:04.661613+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.00535","citing_title":"DREAM-S: Speculative Decoding with Searchable Drafting and Target-Aware Refinement for Multimodal Generation","ref_index":80,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KV237ZSZJIXU5YUMREBM2AV4OJ","json":"https://pith.science/pith/KV237ZSZJIXU5YUMREBM2AV4OJ.json","graph_json":"https://pith.science/api/pith-number/KV237ZSZJIXU5YUMREBM2AV4OJ/graph.json","events_json":"https://pith.science/api/pith-number/KV237ZSZJIXU5YUMREBM2AV4OJ/events.json","paper":"https://pith.science/paper/KV237ZSZ"},"agent_actions":{"view_html":"https://pith.science/pith/KV237ZSZJIXU5YUMREBM2AV4OJ","download_json":"https://pith.science/pith/KV237ZSZJIXU5YUMREBM2AV4OJ.json","view_paper":"https://pith.science/paper/KV237ZSZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.11055&json=true","fetch_graph":"https://pith.science/api/pith-number/KV237ZSZJIXU5YUMREBM2AV4OJ/graph.json","fetch_events":"https://pith.science/api/pith-number/KV237ZSZJIXU5YUMREBM2AV4OJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KV237ZSZJIXU5YUMREBM2AV4OJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KV237ZSZJIXU5YUMREBM2AV4OJ/action/storage_attestation","attest_author":"https://pith.science/pith/KV237ZSZJIXU5YUMREBM2AV4OJ/action/author_attestation","sign_citation":"https://pith.science/pith/KV237ZSZJIXU5YUMREBM2AV4OJ/action/citation_signature","submit_replication":"https://pith.science/pith/KV237ZSZJIXU5YUMREBM2AV4OJ/action/replication_record"}},"created_at":"2026-07-05T11:16:04.661613+00:00","updated_at":"2026-07-05T11:16:04.661613+00:00"}