{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:AAASTQIJNBZN4IWUVH7KAX5QI6","short_pith_number":"pith:AAASTQIJ","schema_version":"1.0","canonical_sha256":"000129c1096872de22d4a9fea05fb047988b0921b5c7c1b4c652866d1aaaa942","source":{"kind":"arxiv","id":"2404.07979","version":2},"attestation_state":"computed","paper":{"title":"LLoCO: Learning Long Contexts Offline","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Joseph E. Gonzalez, Kurt Keutzer, Raluca Ada Popa, Shishir Patil, Sijun Tan, Tianjun Zhang, Xiuyu Li, Ziyang Wu","submitted_at":"2024-04-11T17:57:22Z","abstract_excerpt":"Processing long contexts remains a challenge for large language models (LLMs) due to the quadratic computational and memory overhead of the self-attention mechanism and the substantial KV cache sizes during generation. We propose LLoCO, a novel approach to address this problem by learning contexts offline through context compression and in-domain parameter-efficient finetuning with LoRA. Our method enables an LLM to create a concise representation of the original context and efficiently retrieve relevant information to answer questions accurately. Our approach extends the effective context win"},"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":"2404.07979","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-11T17:57:22Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"21550a6bcf6d6ee9f360f979dd88f7b86d5e2df423cab2d0587008baf7baa745","abstract_canon_sha256":"2a8de25ed19abf0ebd4aa4958dbaa951d34fd1f2d40363312f45d8516270ed66"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:21:42.775927Z","signature_b64":"ggK3KK/Cxvysdl7bmbuG0DK6mmqcP5frZLj5xNNQ299xIzi2JRTyOLCDm/omghMqEqxZLqw9Z0KNTBUEPLt6AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"000129c1096872de22d4a9fea05fb047988b0921b5c7c1b4c652866d1aaaa942","last_reissued_at":"2026-07-05T09:21:42.775426Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:21:42.775426Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLoCO: Learning Long Contexts Offline","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Joseph E. Gonzalez, Kurt Keutzer, Raluca Ada Popa, Shishir Patil, Sijun Tan, Tianjun Zhang, Xiuyu Li, Ziyang Wu","submitted_at":"2024-04-11T17:57:22Z","abstract_excerpt":"Processing long contexts remains a challenge for large language models (LLMs) due to the quadratic computational and memory overhead of the self-attention mechanism and the substantial KV cache sizes during generation. We propose LLoCO, a novel approach to address this problem by learning contexts offline through context compression and in-domain parameter-efficient finetuning with LoRA. Our method enables an LLM to create a concise representation of the original context and efficiently retrieve relevant information to answer questions accurately. Our approach extends the effective context win"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.07979","kind":"arxiv","version":2},"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/2404.07979/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":"2404.07979","created_at":"2026-07-05T09:21:42.775488+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.07979v2","created_at":"2026-07-05T09:21:42.775488+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.07979","created_at":"2026-07-05T09:21:42.775488+00:00"},{"alias_kind":"pith_short_12","alias_value":"AAASTQIJNBZN","created_at":"2026-07-05T09:21:42.775488+00:00"},{"alias_kind":"pith_short_16","alias_value":"AAASTQIJNBZN4IWU","created_at":"2026-07-05T09:21:42.775488+00:00"},{"alias_kind":"pith_short_8","alias_value":"AAASTQIJ","created_at":"2026-07-05T09:21:42.775488+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03979","citing_title":"Language Models Need Sleep: Learning to Self-Modify and Consolidate Memories","ref_index":93,"is_internal_anchor":false},{"citing_arxiv_id":"2409.06679","citing_title":"E2LLM: Encoder Elongated Large Language Models for Long-Context Understanding and Reasoning","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2403.14608","citing_title":"Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey","ref_index":179,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AAASTQIJNBZN4IWUVH7KAX5QI6","json":"https://pith.science/pith/AAASTQIJNBZN4IWUVH7KAX5QI6.json","graph_json":"https://pith.science/api/pith-number/AAASTQIJNBZN4IWUVH7KAX5QI6/graph.json","events_json":"https://pith.science/api/pith-number/AAASTQIJNBZN4IWUVH7KAX5QI6/events.json","paper":"https://pith.science/paper/AAASTQIJ"},"agent_actions":{"view_html":"https://pith.science/pith/AAASTQIJNBZN4IWUVH7KAX5QI6","download_json":"https://pith.science/pith/AAASTQIJNBZN4IWUVH7KAX5QI6.json","view_paper":"https://pith.science/paper/AAASTQIJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.07979&json=true","fetch_graph":"https://pith.science/api/pith-number/AAASTQIJNBZN4IWUVH7KAX5QI6/graph.json","fetch_events":"https://pith.science/api/pith-number/AAASTQIJNBZN4IWUVH7KAX5QI6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AAASTQIJNBZN4IWUVH7KAX5QI6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AAASTQIJNBZN4IWUVH7KAX5QI6/action/storage_attestation","attest_author":"https://pith.science/pith/AAASTQIJNBZN4IWUVH7KAX5QI6/action/author_attestation","sign_citation":"https://pith.science/pith/AAASTQIJNBZN4IWUVH7KAX5QI6/action/citation_signature","submit_replication":"https://pith.science/pith/AAASTQIJNBZN4IWUVH7KAX5QI6/action/replication_record"}},"created_at":"2026-07-05T09:21:42.775488+00:00","updated_at":"2026-07-05T09:21:42.775488+00:00"}