{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:5WIFXJ7XY7EAG5WVVMVTRBCJAL","short_pith_number":"pith:5WIFXJ7X","schema_version":"1.0","canonical_sha256":"ed905ba7f7c7c80376d5ab2b38844902e424b134145945c92076a62ffd2edb67","source":{"kind":"arxiv","id":"2208.04347","version":1},"attestation_state":"computed","paper":{"title":"Investigating Efficiently Extending Transformers for Long Input Summarization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jason Phang, Peter J. Liu, Yao Zhao","submitted_at":"2022-08-08T18:10:58Z","abstract_excerpt":"While large pretrained Transformer models have proven highly capable at tackling natural language tasks, handling long sequence inputs continues to be a significant challenge. One such task is long input summarization, where inputs are longer than the maximum input context of most pretrained models. Through an extensive set of experiments, we investigate what model architectural changes and pretraining paradigms can most efficiently adapt a pretrained Transformer for long input summarization. We find that a staggered, block-local Transformer with global encoder tokens strikes a good balance of"},"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":"2208.04347","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-08-08T18:10:58Z","cross_cats_sorted":[],"title_canon_sha256":"e24881fa6a7af6c330cd87bf46639b6889155bbcef7cefb65ff8d8e5cd2e7db0","abstract_canon_sha256":"2f42bdfc2acb33948c335a8dbcaf68d0fa3646e27c9f8434a6f2e54e11cd4de6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:46:55.708113Z","signature_b64":"7ukEXoKK9ew81GZIn5/sY208RJtCG9liLfRYx04AN8xW69YXG9pwxi6UavJS4WNgso2lA3AWvDsEjMjSeEqcDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ed905ba7f7c7c80376d5ab2b38844902e424b134145945c92076a62ffd2edb67","last_reissued_at":"2026-07-05T04:46:55.707685Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:46:55.707685Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Investigating Efficiently Extending Transformers for Long Input Summarization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jason Phang, Peter J. Liu, Yao Zhao","submitted_at":"2022-08-08T18:10:58Z","abstract_excerpt":"While large pretrained Transformer models have proven highly capable at tackling natural language tasks, handling long sequence inputs continues to be a significant challenge. One such task is long input summarization, where inputs are longer than the maximum input context of most pretrained models. Through an extensive set of experiments, we investigate what model architectural changes and pretraining paradigms can most efficiently adapt a pretrained Transformer for long input summarization. We find that a staggered, block-local Transformer with global encoder tokens strikes a good balance of"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2208.04347","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/2208.04347/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":"2208.04347","created_at":"2026-07-05T04:46:55.707750+00:00"},{"alias_kind":"arxiv_version","alias_value":"2208.04347v1","created_at":"2026-07-05T04:46:55.707750+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2208.04347","created_at":"2026-07-05T04:46:55.707750+00:00"},{"alias_kind":"pith_short_12","alias_value":"5WIFXJ7XY7EA","created_at":"2026-07-05T04:46:55.707750+00:00"},{"alias_kind":"pith_short_16","alias_value":"5WIFXJ7XY7EAG5WV","created_at":"2026-07-05T04:46:55.707750+00:00"},{"alias_kind":"pith_short_8","alias_value":"5WIFXJ7X","created_at":"2026-07-05T04:46:55.707750+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2507.06803","citing_title":"Text to model via SysML: Automated generation of dynamical system computational models from unstructured natural language text via enhanced System Modeling Language diagrams","ref_index":52,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5WIFXJ7XY7EAG5WVVMVTRBCJAL","json":"https://pith.science/pith/5WIFXJ7XY7EAG5WVVMVTRBCJAL.json","graph_json":"https://pith.science/api/pith-number/5WIFXJ7XY7EAG5WVVMVTRBCJAL/graph.json","events_json":"https://pith.science/api/pith-number/5WIFXJ7XY7EAG5WVVMVTRBCJAL/events.json","paper":"https://pith.science/paper/5WIFXJ7X"},"agent_actions":{"view_html":"https://pith.science/pith/5WIFXJ7XY7EAG5WVVMVTRBCJAL","download_json":"https://pith.science/pith/5WIFXJ7XY7EAG5WVVMVTRBCJAL.json","view_paper":"https://pith.science/paper/5WIFXJ7X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2208.04347&json=true","fetch_graph":"https://pith.science/api/pith-number/5WIFXJ7XY7EAG5WVVMVTRBCJAL/graph.json","fetch_events":"https://pith.science/api/pith-number/5WIFXJ7XY7EAG5WVVMVTRBCJAL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5WIFXJ7XY7EAG5WVVMVTRBCJAL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5WIFXJ7XY7EAG5WVVMVTRBCJAL/action/storage_attestation","attest_author":"https://pith.science/pith/5WIFXJ7XY7EAG5WVVMVTRBCJAL/action/author_attestation","sign_citation":"https://pith.science/pith/5WIFXJ7XY7EAG5WVVMVTRBCJAL/action/citation_signature","submit_replication":"https://pith.science/pith/5WIFXJ7XY7EAG5WVVMVTRBCJAL/action/replication_record"}},"created_at":"2026-07-05T04:46:55.707750+00:00","updated_at":"2026-07-05T04:46:55.707750+00:00"}