{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:DVI3NDZGGRJCCUDHPUR6TVJD6V","short_pith_number":"pith:DVI3NDZG","schema_version":"1.0","canonical_sha256":"1d51b68f2634522150677d23e9d523f555abc4b0295b8a19bb0cf2951036da86","source":{"kind":"arxiv","id":"2408.10277","version":1},"attestation_state":"computed","paper":{"title":"Increasing transformer token length with a Maximum Entropy Principle Method","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"R. I. Cukier","submitted_at":"2024-08-17T15:47:39Z","abstract_excerpt":"Transformers suffer from the computational overhead of their quadratic dependence on the length of sequences processed. We present three methods, all adding an intermediate step between training and inference/generation, which extend the autoregressive length of transformers. All rely on a Maximum Entropy Principle (MEP) whereby entropy is maximized in the presence of suitable constraints, accounted for by use of Lagrange Multipliers. These constraint methods extend the autoregressive character from T to 2T tokens in a linear-with-T fashion. There is overhead associated with this added step, 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":"2408.10277","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-17T15:47:39Z","cross_cats_sorted":[],"title_canon_sha256":"e0c42c0a33ee8436d25b8ab97aea4d3ae978e1b6e75876126f59f3d13dc5959f","abstract_canon_sha256":"d1a18841ff501bbcb977765614a7b1105f12fd779135812987b4569c48946892"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:57:09.108738Z","signature_b64":"95bDIUV/ri/OvD0Iqet2eCWjkjKFp1/HlWGIuZaD9yA2CHOg0uWFVm1+RoXnrfKNRe4AyvRMrqw7c4N2Pyy0Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1d51b68f2634522150677d23e9d523f555abc4b0295b8a19bb0cf2951036da86","last_reissued_at":"2026-07-05T08:57:09.108305Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:57:09.108305Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Increasing transformer token length with a Maximum Entropy Principle Method","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"R. I. Cukier","submitted_at":"2024-08-17T15:47:39Z","abstract_excerpt":"Transformers suffer from the computational overhead of their quadratic dependence on the length of sequences processed. We present three methods, all adding an intermediate step between training and inference/generation, which extend the autoregressive length of transformers. All rely on a Maximum Entropy Principle (MEP) whereby entropy is maximized in the presence of suitable constraints, accounted for by use of Lagrange Multipliers. These constraint methods extend the autoregressive character from T to 2T tokens in a linear-with-T fashion. There is overhead associated with this added step, b"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.10277","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/2408.10277/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":"2408.10277","created_at":"2026-07-05T08:57:09.108371+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.10277v1","created_at":"2026-07-05T08:57:09.108371+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.10277","created_at":"2026-07-05T08:57:09.108371+00:00"},{"alias_kind":"pith_short_12","alias_value":"DVI3NDZGGRJC","created_at":"2026-07-05T08:57:09.108371+00:00"},{"alias_kind":"pith_short_16","alias_value":"DVI3NDZGGRJCCUDH","created_at":"2026-07-05T08:57:09.108371+00:00"},{"alias_kind":"pith_short_8","alias_value":"DVI3NDZG","created_at":"2026-07-05T08:57:09.108371+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.07680","citing_title":"SpecRouter: Adaptive Routing for Multi-Level Speculative Decoding in Large Language Models","ref_index":13,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DVI3NDZGGRJCCUDHPUR6TVJD6V","json":"https://pith.science/pith/DVI3NDZGGRJCCUDHPUR6TVJD6V.json","graph_json":"https://pith.science/api/pith-number/DVI3NDZGGRJCCUDHPUR6TVJD6V/graph.json","events_json":"https://pith.science/api/pith-number/DVI3NDZGGRJCCUDHPUR6TVJD6V/events.json","paper":"https://pith.science/paper/DVI3NDZG"},"agent_actions":{"view_html":"https://pith.science/pith/DVI3NDZGGRJCCUDHPUR6TVJD6V","download_json":"https://pith.science/pith/DVI3NDZGGRJCCUDHPUR6TVJD6V.json","view_paper":"https://pith.science/paper/DVI3NDZG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.10277&json=true","fetch_graph":"https://pith.science/api/pith-number/DVI3NDZGGRJCCUDHPUR6TVJD6V/graph.json","fetch_events":"https://pith.science/api/pith-number/DVI3NDZGGRJCCUDHPUR6TVJD6V/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DVI3NDZGGRJCCUDHPUR6TVJD6V/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DVI3NDZGGRJCCUDHPUR6TVJD6V/action/storage_attestation","attest_author":"https://pith.science/pith/DVI3NDZGGRJCCUDHPUR6TVJD6V/action/author_attestation","sign_citation":"https://pith.science/pith/DVI3NDZGGRJCCUDHPUR6TVJD6V/action/citation_signature","submit_replication":"https://pith.science/pith/DVI3NDZGGRJCCUDHPUR6TVJD6V/action/replication_record"}},"created_at":"2026-07-05T08:57:09.108371+00:00","updated_at":"2026-07-05T08:57:09.108371+00:00"}