{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:2FKXDTJN4QH2CUMPFTPRLZONGA","short_pith_number":"pith:2FKXDTJN","schema_version":"1.0","canonical_sha256":"d15571cd2de40fa1518f2cdf15e5cd3000603d54614e952f8db25e6af80893f6","source":{"kind":"arxiv","id":"2501.12766","version":2},"attestation_state":"computed","paper":{"title":"NExtLong: Toward Effective Long-Context Training without Long Documents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chaochen Gao, Debing Zhang, Songlin Hu, Xing Wu, Zijia Lin","submitted_at":"2025-01-22T10:01:54Z","abstract_excerpt":"Large language models (LLMs) with extended context windows have made significant strides yet remain a challenge due to the scarcity of long documents. Existing methods tend to synthesize long-context data but lack a clear mechanism to reinforce the long-range dependency modeling. To address this limitation, we propose NExtLong, a novel framework for synthesizing long-context data through Negative document Extension. NExtLong decomposes a document into multiple meta-chunks and extends the context by interleaving hard negative distractors retrieved from pretraining corpora. This approach compels"},"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":"2501.12766","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-22T10:01:54Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3010e1a95c4eaa9aceb23a0d89ab00b1e940bdf3531a545b2f9a12734358cc5b","abstract_canon_sha256":"b22e61c82a88efd193681aeff09698f1ab8b5df5655f49c7f37e7ccd8511335d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:09:06.912240Z","signature_b64":"iyXSAX/tXkyViKW42O+tC2kKvs1Gz+60qdZunF+ligeXfNq7fdKws9O6+lWfy6eEUZGOv7QLtKMaDxWEg112BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d15571cd2de40fa1518f2cdf15e5cd3000603d54614e952f8db25e6af80893f6","last_reissued_at":"2026-07-05T11:09:06.911762Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:09:06.911762Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"NExtLong: Toward Effective Long-Context Training without Long Documents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chaochen Gao, Debing Zhang, Songlin Hu, Xing Wu, Zijia Lin","submitted_at":"2025-01-22T10:01:54Z","abstract_excerpt":"Large language models (LLMs) with extended context windows have made significant strides yet remain a challenge due to the scarcity of long documents. Existing methods tend to synthesize long-context data but lack a clear mechanism to reinforce the long-range dependency modeling. To address this limitation, we propose NExtLong, a novel framework for synthesizing long-context data through Negative document Extension. NExtLong decomposes a document into multiple meta-chunks and extends the context by interleaving hard negative distractors retrieved from pretraining corpora. This approach compels"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.12766","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/2501.12766/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":"2501.12766","created_at":"2026-07-05T11:09:06.911819+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.12766v2","created_at":"2026-07-05T11:09:06.911819+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.12766","created_at":"2026-07-05T11:09:06.911819+00:00"},{"alias_kind":"pith_short_12","alias_value":"2FKXDTJN4QH2","created_at":"2026-07-05T11:09:06.911819+00:00"},{"alias_kind":"pith_short_16","alias_value":"2FKXDTJN4QH2CUMP","created_at":"2026-07-05T11:09:06.911819+00:00"},{"alias_kind":"pith_short_8","alias_value":"2FKXDTJN","created_at":"2026-07-05T11:09:06.911819+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2507.02259","citing_title":"MemAgent: Reshaping Long-Context LLM with Multi-Conv RL-based Memory Agent","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2502.17419","citing_title":"From System 1 to System 2: A Survey of Reasoning Large Language Models","ref_index":253,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2FKXDTJN4QH2CUMPFTPRLZONGA","json":"https://pith.science/pith/2FKXDTJN4QH2CUMPFTPRLZONGA.json","graph_json":"https://pith.science/api/pith-number/2FKXDTJN4QH2CUMPFTPRLZONGA/graph.json","events_json":"https://pith.science/api/pith-number/2FKXDTJN4QH2CUMPFTPRLZONGA/events.json","paper":"https://pith.science/paper/2FKXDTJN"},"agent_actions":{"view_html":"https://pith.science/pith/2FKXDTJN4QH2CUMPFTPRLZONGA","download_json":"https://pith.science/pith/2FKXDTJN4QH2CUMPFTPRLZONGA.json","view_paper":"https://pith.science/paper/2FKXDTJN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.12766&json=true","fetch_graph":"https://pith.science/api/pith-number/2FKXDTJN4QH2CUMPFTPRLZONGA/graph.json","fetch_events":"https://pith.science/api/pith-number/2FKXDTJN4QH2CUMPFTPRLZONGA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2FKXDTJN4QH2CUMPFTPRLZONGA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2FKXDTJN4QH2CUMPFTPRLZONGA/action/storage_attestation","attest_author":"https://pith.science/pith/2FKXDTJN4QH2CUMPFTPRLZONGA/action/author_attestation","sign_citation":"https://pith.science/pith/2FKXDTJN4QH2CUMPFTPRLZONGA/action/citation_signature","submit_replication":"https://pith.science/pith/2FKXDTJN4QH2CUMPFTPRLZONGA/action/replication_record"}},"created_at":"2026-07-05T11:09:06.911819+00:00","updated_at":"2026-07-05T11:09:06.911819+00:00"}