{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:IK3XZCSNEPIYKEN4SNTQKVULYJ","short_pith_number":"pith:IK3XZCSN","schema_version":"1.0","canonical_sha256":"42b77c8a4d23d18511bc936705568bc27376572b7a0cfba9633de8718b76cef1","source":{"kind":"arxiv","id":"2405.00888","version":1},"attestation_state":"computed","paper":{"title":"DynaMo: Accelerating Language Model Inference with Dynamic Multi-Token Sampling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chi-Heng Lin, Hongxia Jin, Niraj K. Jha, Shikhar Tuli, Yen-Chang Hsu, Yilin Shen","submitted_at":"2024-05-01T22:17:57Z","abstract_excerpt":"Traditional language models operate autoregressively, i.e., they predict one token at a time. Rapid explosion in model sizes has resulted in high inference times. In this work, we propose DynaMo, a suite of multi-token prediction language models that reduce net inference times. Our models $\\textit{dynamically}$ predict multiple tokens based on their confidence in the predicted joint probability distribution. We propose a lightweight technique to train these models, leveraging the weights of traditional autoregressive counterparts. Moreover, we propose novel ways to enhance the estimated joint "},"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":"2405.00888","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-01T22:17:57Z","cross_cats_sorted":[],"title_canon_sha256":"ccb2978e9fd189678cde70d48bb6b20e4e94a605238be600399f82e04fa744ec","abstract_canon_sha256":"d61309e9d443248d371b5fbb86d9f0db81681ae1570dd89a98e9b2da121f57b4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:14:27.611373Z","signature_b64":"KF+/pcjcHQOKHEO6rlHqfNH/iLW4N6L/PkilQ0Bv2XJR42sFs1TCa+ATy3UB7Mh8j39yenpeBEyea0orCw72CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"42b77c8a4d23d18511bc936705568bc27376572b7a0cfba9633de8718b76cef1","last_reissued_at":"2026-07-05T08:14:27.610848Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:14:27.610848Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DynaMo: Accelerating Language Model Inference with Dynamic Multi-Token Sampling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chi-Heng Lin, Hongxia Jin, Niraj K. Jha, Shikhar Tuli, Yen-Chang Hsu, Yilin Shen","submitted_at":"2024-05-01T22:17:57Z","abstract_excerpt":"Traditional language models operate autoregressively, i.e., they predict one token at a time. Rapid explosion in model sizes has resulted in high inference times. In this work, we propose DynaMo, a suite of multi-token prediction language models that reduce net inference times. Our models $\\textit{dynamically}$ predict multiple tokens based on their confidence in the predicted joint probability distribution. We propose a lightweight technique to train these models, leveraging the weights of traditional autoregressive counterparts. Moreover, we propose novel ways to enhance the estimated joint "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.00888","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/2405.00888/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":"2405.00888","created_at":"2026-07-05T08:14:27.610909+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.00888v1","created_at":"2026-07-05T08:14:27.610909+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.00888","created_at":"2026-07-05T08:14:27.610909+00:00"},{"alias_kind":"pith_short_12","alias_value":"IK3XZCSNEPIY","created_at":"2026-07-05T08:14:27.610909+00:00"},{"alias_kind":"pith_short_16","alias_value":"IK3XZCSNEPIYKEN4","created_at":"2026-07-05T08:14:27.610909+00:00"},{"alias_kind":"pith_short_8","alias_value":"IK3XZCSN","created_at":"2026-07-05T08:14:27.610909+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.09955","citing_title":"DeepSeek: Paradigm Shifts and Technical Evolution in Large AI Models","ref_index":114,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IK3XZCSNEPIYKEN4SNTQKVULYJ","json":"https://pith.science/pith/IK3XZCSNEPIYKEN4SNTQKVULYJ.json","graph_json":"https://pith.science/api/pith-number/IK3XZCSNEPIYKEN4SNTQKVULYJ/graph.json","events_json":"https://pith.science/api/pith-number/IK3XZCSNEPIYKEN4SNTQKVULYJ/events.json","paper":"https://pith.science/paper/IK3XZCSN"},"agent_actions":{"view_html":"https://pith.science/pith/IK3XZCSNEPIYKEN4SNTQKVULYJ","download_json":"https://pith.science/pith/IK3XZCSNEPIYKEN4SNTQKVULYJ.json","view_paper":"https://pith.science/paper/IK3XZCSN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.00888&json=true","fetch_graph":"https://pith.science/api/pith-number/IK3XZCSNEPIYKEN4SNTQKVULYJ/graph.json","fetch_events":"https://pith.science/api/pith-number/IK3XZCSNEPIYKEN4SNTQKVULYJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IK3XZCSNEPIYKEN4SNTQKVULYJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IK3XZCSNEPIYKEN4SNTQKVULYJ/action/storage_attestation","attest_author":"https://pith.science/pith/IK3XZCSNEPIYKEN4SNTQKVULYJ/action/author_attestation","sign_citation":"https://pith.science/pith/IK3XZCSNEPIYKEN4SNTQKVULYJ/action/citation_signature","submit_replication":"https://pith.science/pith/IK3XZCSNEPIYKEN4SNTQKVULYJ/action/replication_record"}},"created_at":"2026-07-05T08:14:27.610909+00:00","updated_at":"2026-07-05T08:14:27.610909+00:00"}