{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:GTI7SIFKYAHQZFIHRM7ELMSOZ5","short_pith_number":"pith:GTI7SIFK","schema_version":"1.0","canonical_sha256":"34d1f920aac00f0c95078b3e45b24ecf721c5293276a96c0c7e36ef768fdc031","source":{"kind":"arxiv","id":"2305.20019","version":1},"attestation_state":"computed","paper":{"title":"Monotonic Location Attention for Length Generalization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Cornelia Caragea, Jishnu Ray Chowdhury","submitted_at":"2023-05-31T16:48:06Z","abstract_excerpt":"We explore different ways to utilize position-based cross-attention in seq2seq networks to enable length generalization in algorithmic tasks. We show that a simple approach of interpolating the original and reversed encoded representations combined with relative attention allows near-perfect length generalization for both forward and reverse lookup tasks or copy tasks that had been generally hard to tackle. We also devise harder diagnostic tasks where the relative distance of the ideal attention position varies with timestep. In such settings, the simple interpolation trick with relative atten"},"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":"2305.20019","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-05-31T16:48:06Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"aa729594b1c9a8de1b64bd923ca17413dd9fb2bfe7a011a5cdf7a4bd0dafccab","abstract_canon_sha256":"f60c44d7847eb995fd72ca79692ff0c915f78f6743a7190fb4caae46692fa332"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:16:05.958052Z","signature_b64":"RVVLfxYx5HCrMx+Ry+62y1/jVvPR8ftLEjvOcGV4QNaibDmxlmSbIQtzoUIjwf53BaqSupIruDvKW7ade0X4Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"34d1f920aac00f0c95078b3e45b24ecf721c5293276a96c0c7e36ef768fdc031","last_reissued_at":"2026-07-05T06:16:05.957592Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:16:05.957592Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Monotonic Location Attention for Length Generalization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Cornelia Caragea, Jishnu Ray Chowdhury","submitted_at":"2023-05-31T16:48:06Z","abstract_excerpt":"We explore different ways to utilize position-based cross-attention in seq2seq networks to enable length generalization in algorithmic tasks. We show that a simple approach of interpolating the original and reversed encoded representations combined with relative attention allows near-perfect length generalization for both forward and reverse lookup tasks or copy tasks that had been generally hard to tackle. We also devise harder diagnostic tasks where the relative distance of the ideal attention position varies with timestep. In such settings, the simple interpolation trick with relative atten"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.20019","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/2305.20019/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":"2305.20019","created_at":"2026-07-05T06:16:05.957647+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.20019v1","created_at":"2026-07-05T06:16:05.957647+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.20019","created_at":"2026-07-05T06:16:05.957647+00:00"},{"alias_kind":"pith_short_12","alias_value":"GTI7SIFKYAHQ","created_at":"2026-07-05T06:16:05.957647+00:00"},{"alias_kind":"pith_short_16","alias_value":"GTI7SIFKYAHQZFIH","created_at":"2026-07-05T06:16:05.957647+00:00"},{"alias_kind":"pith_short_8","alias_value":"GTI7SIFK","created_at":"2026-07-05T06:16:05.957647+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2509.05218","citing_title":"HoPE: Hyperbolic Rotary Positional Encoding for Stable Long-Range Dependency Modeling in Large Language Models","ref_index":12,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GTI7SIFKYAHQZFIHRM7ELMSOZ5","json":"https://pith.science/pith/GTI7SIFKYAHQZFIHRM7ELMSOZ5.json","graph_json":"https://pith.science/api/pith-number/GTI7SIFKYAHQZFIHRM7ELMSOZ5/graph.json","events_json":"https://pith.science/api/pith-number/GTI7SIFKYAHQZFIHRM7ELMSOZ5/events.json","paper":"https://pith.science/paper/GTI7SIFK"},"agent_actions":{"view_html":"https://pith.science/pith/GTI7SIFKYAHQZFIHRM7ELMSOZ5","download_json":"https://pith.science/pith/GTI7SIFKYAHQZFIHRM7ELMSOZ5.json","view_paper":"https://pith.science/paper/GTI7SIFK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.20019&json=true","fetch_graph":"https://pith.science/api/pith-number/GTI7SIFKYAHQZFIHRM7ELMSOZ5/graph.json","fetch_events":"https://pith.science/api/pith-number/GTI7SIFKYAHQZFIHRM7ELMSOZ5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GTI7SIFKYAHQZFIHRM7ELMSOZ5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GTI7SIFKYAHQZFIHRM7ELMSOZ5/action/storage_attestation","attest_author":"https://pith.science/pith/GTI7SIFKYAHQZFIHRM7ELMSOZ5/action/author_attestation","sign_citation":"https://pith.science/pith/GTI7SIFKYAHQZFIHRM7ELMSOZ5/action/citation_signature","submit_replication":"https://pith.science/pith/GTI7SIFKYAHQZFIHRM7ELMSOZ5/action/replication_record"}},"created_at":"2026-07-05T06:16:05.957647+00:00","updated_at":"2026-07-05T06:16:05.957647+00:00"}