{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:DXTS3W5XBZ4VQCXJQWQ7AXK6SW","short_pith_number":"pith:DXTS3W5X","schema_version":"1.0","canonical_sha256":"1de72ddbb70e79580ae985a1f05d5e95a0117a3875b43fe3d18f128492395475","source":{"kind":"arxiv","id":"2109.07276","version":1},"attestation_state":"computed","paper":{"title":"Sequence Length is a Domain: Length-based Overfitting in Transformer Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Du\\v{s}an Vari\\v{s}, Ond\\v{r}ej Bojar","submitted_at":"2021-09-15T13:25:19Z","abstract_excerpt":"Transformer-based sequence-to-sequence architectures, while achieving state-of-the-art results on a large number of NLP tasks, can still suffer from overfitting during training. In practice, this is usually countered either by applying regularization methods (e.g. dropout, L2-regularization) or by providing huge amounts of training data. Additionally, Transformer and other architectures are known to struggle when generating very long sequences. For example, in machine translation, the neural-based systems perform worse on very long sequences when compared to the preceding phrase-based translat"},"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":"2109.07276","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-09-15T13:25:19Z","cross_cats_sorted":[],"title_canon_sha256":"bf21d8d5c5395d102c5b34fb26b2721a2f234d6a2d43f7a8c5ba1b215d4e668a","abstract_canon_sha256":"9835cad86248a9c0d39a3c211bf55003ca45cfd768a189df4dc6b6105f0e321f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:44:55.670809Z","signature_b64":"XZFJ58WevjDrIQWvEt2tsOPGY3+A+fszUQKxoayMfbPTfn3eC2g6F+9cyp9Sk59lyxTQccsty/iYcORSVCppAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1de72ddbb70e79580ae985a1f05d5e95a0117a3875b43fe3d18f128492395475","last_reissued_at":"2026-07-05T03:44:55.670322Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:44:55.670322Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sequence Length is a Domain: Length-based Overfitting in Transformer Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Du\\v{s}an Vari\\v{s}, Ond\\v{r}ej Bojar","submitted_at":"2021-09-15T13:25:19Z","abstract_excerpt":"Transformer-based sequence-to-sequence architectures, while achieving state-of-the-art results on a large number of NLP tasks, can still suffer from overfitting during training. In practice, this is usually countered either by applying regularization methods (e.g. dropout, L2-regularization) or by providing huge amounts of training data. Additionally, Transformer and other architectures are known to struggle when generating very long sequences. For example, in machine translation, the neural-based systems perform worse on very long sequences when compared to the preceding phrase-based translat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.07276","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/2109.07276/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":"2109.07276","created_at":"2026-07-05T03:44:55.670378+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.07276v1","created_at":"2026-07-05T03:44:55.670378+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.07276","created_at":"2026-07-05T03:44:55.670378+00:00"},{"alias_kind":"pith_short_12","alias_value":"DXTS3W5XBZ4V","created_at":"2026-07-05T03:44:55.670378+00:00"},{"alias_kind":"pith_short_16","alias_value":"DXTS3W5XBZ4VQCXJ","created_at":"2026-07-05T03:44:55.670378+00:00"},{"alias_kind":"pith_short_8","alias_value":"DXTS3W5X","created_at":"2026-07-05T03:44:55.670378+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2306.14048","citing_title":"H$_2$O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models","ref_index":84,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DXTS3W5XBZ4VQCXJQWQ7AXK6SW","json":"https://pith.science/pith/DXTS3W5XBZ4VQCXJQWQ7AXK6SW.json","graph_json":"https://pith.science/api/pith-number/DXTS3W5XBZ4VQCXJQWQ7AXK6SW/graph.json","events_json":"https://pith.science/api/pith-number/DXTS3W5XBZ4VQCXJQWQ7AXK6SW/events.json","paper":"https://pith.science/paper/DXTS3W5X"},"agent_actions":{"view_html":"https://pith.science/pith/DXTS3W5XBZ4VQCXJQWQ7AXK6SW","download_json":"https://pith.science/pith/DXTS3W5XBZ4VQCXJQWQ7AXK6SW.json","view_paper":"https://pith.science/paper/DXTS3W5X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.07276&json=true","fetch_graph":"https://pith.science/api/pith-number/DXTS3W5XBZ4VQCXJQWQ7AXK6SW/graph.json","fetch_events":"https://pith.science/api/pith-number/DXTS3W5XBZ4VQCXJQWQ7AXK6SW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DXTS3W5XBZ4VQCXJQWQ7AXK6SW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DXTS3W5XBZ4VQCXJQWQ7AXK6SW/action/storage_attestation","attest_author":"https://pith.science/pith/DXTS3W5XBZ4VQCXJQWQ7AXK6SW/action/author_attestation","sign_citation":"https://pith.science/pith/DXTS3W5XBZ4VQCXJQWQ7AXK6SW/action/citation_signature","submit_replication":"https://pith.science/pith/DXTS3W5XBZ4VQCXJQWQ7AXK6SW/action/replication_record"}},"created_at":"2026-07-05T03:44:55.670378+00:00","updated_at":"2026-07-05T03:44:55.670378+00:00"}