{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:B7OOWUTMNHHDTLPWWIJBKOQEJX","short_pith_number":"pith:B7OOWUTM","schema_version":"1.0","canonical_sha256":"0fdceb526c69ce39adf6b212153a044dd4ebf61bf09ccc12865f357b7a56ab2e","source":{"kind":"arxiv","id":"2503.16071","version":1},"attestation_state":"computed","paper":{"title":"Tuning LLMs by RAG Principles: Towards LLM-native Memory","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.IR"],"primary_cat":"cs.CL","authors_text":"Fangbo Tao, Jiale Wei, Jingbo Shang, Ruochen Liu, Shuchi Wu, Xiang Ying","submitted_at":"2025-03-20T12:04:40Z","abstract_excerpt":"Memory, additional information beyond the training of large language models (LLMs), is crucial to various real-world applications, such as personal assistant. The two mainstream solutions to incorporate memory into the generation process are long-context LLMs and retrieval-augmented generation (RAG). In this paper, we first systematically compare these two types of solutions on three renovated/new datasets and show that (1) long-context solutions, although more expensive, shall be easier to capture the big picture and better answer queries which require considering the memory as a whole; and ("},"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":"2503.16071","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-20T12:04:40Z","cross_cats_sorted":["cs.AI","cs.IR"],"title_canon_sha256":"c57b4856026c6cb43a756eedf297931870a1033cd811270573c80cd2ef5dcaaf","abstract_canon_sha256":"6c19c826047c47b59f6dbc64ebef83d6777d7ffad85214c2bf91107b51d5bff5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:36:05.369364Z","signature_b64":"i/1XKqZXs6xo/7C6hFtu4jI0I9JS6x4ZGS3cUYIhMfVHCiW70QoHyO6h2suWdNIKdfjLp2BRsNVon83T2z1IBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0fdceb526c69ce39adf6b212153a044dd4ebf61bf09ccc12865f357b7a56ab2e","last_reissued_at":"2026-07-05T10:36:05.368637Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:36:05.368637Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Tuning LLMs by RAG Principles: Towards LLM-native Memory","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.IR"],"primary_cat":"cs.CL","authors_text":"Fangbo Tao, Jiale Wei, Jingbo Shang, Ruochen Liu, Shuchi Wu, Xiang Ying","submitted_at":"2025-03-20T12:04:40Z","abstract_excerpt":"Memory, additional information beyond the training of large language models (LLMs), is crucial to various real-world applications, such as personal assistant. The two mainstream solutions to incorporate memory into the generation process are long-context LLMs and retrieval-augmented generation (RAG). In this paper, we first systematically compare these two types of solutions on three renovated/new datasets and show that (1) long-context solutions, although more expensive, shall be easier to capture the big picture and better answer queries which require considering the memory as a whole; and ("},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.16071","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/2503.16071/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":"2503.16071","created_at":"2026-07-05T10:36:05.368727+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.16071v1","created_at":"2026-07-05T10:36:05.368727+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.16071","created_at":"2026-07-05T10:36:05.368727+00:00"},{"alias_kind":"pith_short_12","alias_value":"B7OOWUTMNHHD","created_at":"2026-07-05T10:36:05.368727+00:00"},{"alias_kind":"pith_short_16","alias_value":"B7OOWUTMNHHDTLPW","created_at":"2026-07-05T10:36:05.368727+00:00"},{"alias_kind":"pith_short_8","alias_value":"B7OOWUTM","created_at":"2026-07-05T10:36:05.368727+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.02162","citing_title":"AAFLOW: Scalable Patterns for Agentic AI Workflows","ref_index":64,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/B7OOWUTMNHHDTLPWWIJBKOQEJX","json":"https://pith.science/pith/B7OOWUTMNHHDTLPWWIJBKOQEJX.json","graph_json":"https://pith.science/api/pith-number/B7OOWUTMNHHDTLPWWIJBKOQEJX/graph.json","events_json":"https://pith.science/api/pith-number/B7OOWUTMNHHDTLPWWIJBKOQEJX/events.json","paper":"https://pith.science/paper/B7OOWUTM"},"agent_actions":{"view_html":"https://pith.science/pith/B7OOWUTMNHHDTLPWWIJBKOQEJX","download_json":"https://pith.science/pith/B7OOWUTMNHHDTLPWWIJBKOQEJX.json","view_paper":"https://pith.science/paper/B7OOWUTM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.16071&json=true","fetch_graph":"https://pith.science/api/pith-number/B7OOWUTMNHHDTLPWWIJBKOQEJX/graph.json","fetch_events":"https://pith.science/api/pith-number/B7OOWUTMNHHDTLPWWIJBKOQEJX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/B7OOWUTMNHHDTLPWWIJBKOQEJX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/B7OOWUTMNHHDTLPWWIJBKOQEJX/action/storage_attestation","attest_author":"https://pith.science/pith/B7OOWUTMNHHDTLPWWIJBKOQEJX/action/author_attestation","sign_citation":"https://pith.science/pith/B7OOWUTMNHHDTLPWWIJBKOQEJX/action/citation_signature","submit_replication":"https://pith.science/pith/B7OOWUTMNHHDTLPWWIJBKOQEJX/action/replication_record"}},"created_at":"2026-07-05T10:36:05.368727+00:00","updated_at":"2026-07-05T10:36:05.368727+00:00"}