{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:2HAJJOONRDQXZSJR263SAMJCZK","short_pith_number":"pith:2HAJJOON","schema_version":"1.0","canonical_sha256":"d1c094b9cd88e17cc931d7b7203122caac2c55cde150d620650afaa191219be2","source":{"kind":"arxiv","id":"2312.11193","version":10},"attestation_state":"computed","paper":{"title":"Training With \"Paraphrasing the Original Text\" Teaches LLM to Better Retrieve in Long-context Tasks","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Yijiong Yu, Yongfeng Huang, Zhe Zhou, Zhixiao Qi","submitted_at":"2023-12-18T13:40:16Z","abstract_excerpt":"As Large Language Models (LLMs) continue to evolve, more are being designed to handle long-context inputs. Despite this advancement, most of them still face challenges in accurately handling long-context tasks, often showing the \"lost in the middle\" issue. We identify that insufficient retrieval capability is one of the important reasons for this issue. To tackle this challenge, we propose a novel approach to design training data for long-context tasks, aiming at augmenting LLMs' proficiency in extracting key information from long context. Specially, we incorporate an additional part named \"pa"},"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":"2312.11193","kind":"arxiv","version":10},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2023-12-18T13:40:16Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"033bd48bf9413ec7f34558f59e31a7c5180b68e00e9a463a6aa8a6b554f46190","abstract_canon_sha256":"4df7988cc931ec887a61bf76b21be50dc964fb2caba99ffdeed38a4cde73a0a0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:47:54.170861Z","signature_b64":"KGCm8dMBhJCxQfRTHSyjcEjAKu7ljZy930T+vKca/OtfvqTzgPKkvnoMlmVNQNzqMKwRioTIvbMtcC5RWZF9DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d1c094b9cd88e17cc931d7b7203122caac2c55cde150d620650afaa191219be2","last_reissued_at":"2026-07-05T09:47:54.170380Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:47:54.170380Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Training With \"Paraphrasing the Original Text\" Teaches LLM to Better Retrieve in Long-context Tasks","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Yijiong Yu, Yongfeng Huang, Zhe Zhou, Zhixiao Qi","submitted_at":"2023-12-18T13:40:16Z","abstract_excerpt":"As Large Language Models (LLMs) continue to evolve, more are being designed to handle long-context inputs. Despite this advancement, most of them still face challenges in accurately handling long-context tasks, often showing the \"lost in the middle\" issue. We identify that insufficient retrieval capability is one of the important reasons for this issue. To tackle this challenge, we propose a novel approach to design training data for long-context tasks, aiming at augmenting LLMs' proficiency in extracting key information from long context. Specially, we incorporate an additional part named \"pa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.11193","kind":"arxiv","version":10},"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/2312.11193/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":"2312.11193","created_at":"2026-07-05T09:47:54.170448+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.11193v10","created_at":"2026-07-05T09:47:54.170448+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.11193","created_at":"2026-07-05T09:47:54.170448+00:00"},{"alias_kind":"pith_short_12","alias_value":"2HAJJOONRDQX","created_at":"2026-07-05T09:47:54.170448+00:00"},{"alias_kind":"pith_short_16","alias_value":"2HAJJOONRDQXZSJR","created_at":"2026-07-05T09:47:54.170448+00:00"},{"alias_kind":"pith_short_8","alias_value":"2HAJJOON","created_at":"2026-07-05T09:47:54.170448+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2403.04652","citing_title":"Yi: Open Foundation Models by 01.AI","ref_index":89,"is_internal_anchor":false},{"citing_arxiv_id":"2412.05271","citing_title":"Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling","ref_index":287,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2HAJJOONRDQXZSJR263SAMJCZK","json":"https://pith.science/pith/2HAJJOONRDQXZSJR263SAMJCZK.json","graph_json":"https://pith.science/api/pith-number/2HAJJOONRDQXZSJR263SAMJCZK/graph.json","events_json":"https://pith.science/api/pith-number/2HAJJOONRDQXZSJR263SAMJCZK/events.json","paper":"https://pith.science/paper/2HAJJOON"},"agent_actions":{"view_html":"https://pith.science/pith/2HAJJOONRDQXZSJR263SAMJCZK","download_json":"https://pith.science/pith/2HAJJOONRDQXZSJR263SAMJCZK.json","view_paper":"https://pith.science/paper/2HAJJOON","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.11193&json=true","fetch_graph":"https://pith.science/api/pith-number/2HAJJOONRDQXZSJR263SAMJCZK/graph.json","fetch_events":"https://pith.science/api/pith-number/2HAJJOONRDQXZSJR263SAMJCZK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2HAJJOONRDQXZSJR263SAMJCZK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2HAJJOONRDQXZSJR263SAMJCZK/action/storage_attestation","attest_author":"https://pith.science/pith/2HAJJOONRDQXZSJR263SAMJCZK/action/author_attestation","sign_citation":"https://pith.science/pith/2HAJJOONRDQXZSJR263SAMJCZK/action/citation_signature","submit_replication":"https://pith.science/pith/2HAJJOONRDQXZSJR263SAMJCZK/action/replication_record"}},"created_at":"2026-07-05T09:47:54.170448+00:00","updated_at":"2026-07-05T09:47:54.170448+00:00"}