{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:2LYOCNG623XRAJDQJQAC5NFGIM","short_pith_number":"pith:2LYOCNG6","schema_version":"1.0","canonical_sha256":"d2f0e134ded6ef1024704c002eb4a6432441fbb69da69d91db8e87509fbbcf1c","source":{"kind":"arxiv","id":"2404.11672","version":3},"attestation_state":"computed","paper":{"title":"MemLLM: Finetuning LLMs to Use An Explicit Read-Write Memory","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Abdullatif K\\\"oksal, Ali Modarressi, Ayyoob Imani, Hinrich Sch\\\"utze, Mohsen Fayyaz","submitted_at":"2024-04-17T18:13:16Z","abstract_excerpt":"While current large language models (LLMs) perform well on many knowledge-related tasks, they are limited by relying on their parameters as an implicit storage mechanism. As a result, they struggle with memorizing rare events and with updating their memory as facts change over time. In addition, the uninterpretable nature of parametric memory makes it challenging to prevent hallucination. Model editing and augmenting LLMs with parameters specialized for memory are only partial solutions. In this paper, we introduce MemLLM, a novel method of enhancing LLMs by integrating a structured and explic"},"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":"2404.11672","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-17T18:13:16Z","cross_cats_sorted":[],"title_canon_sha256":"109ddfa0c4b44e965f39d0f414e57fa9cfa473fff5e45bbf858534b37d7bdcad","abstract_canon_sha256":"3be0a7ef57535a54e64319f974913f3eefbc33a037bc1e87e004dd3637ab182b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:50:29.659130Z","signature_b64":"d9iOeEurSR9s58oTEIkqfen9HYwJ3qyFqtAuOooq9Kz4y3x951/8bmhvWfat9g7LxMDadnDyD/DNXqct85npDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d2f0e134ded6ef1024704c002eb4a6432441fbb69da69d91db8e87509fbbcf1c","last_reissued_at":"2026-07-05T10:50:29.658583Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:50:29.658583Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MemLLM: Finetuning LLMs to Use An Explicit Read-Write Memory","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Abdullatif K\\\"oksal, Ali Modarressi, Ayyoob Imani, Hinrich Sch\\\"utze, Mohsen Fayyaz","submitted_at":"2024-04-17T18:13:16Z","abstract_excerpt":"While current large language models (LLMs) perform well on many knowledge-related tasks, they are limited by relying on their parameters as an implicit storage mechanism. As a result, they struggle with memorizing rare events and with updating their memory as facts change over time. In addition, the uninterpretable nature of parametric memory makes it challenging to prevent hallucination. Model editing and augmenting LLMs with parameters specialized for memory are only partial solutions. In this paper, we introduce MemLLM, a novel method of enhancing LLMs by integrating a structured and explic"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.11672","kind":"arxiv","version":3},"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/2404.11672/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":"2404.11672","created_at":"2026-07-05T10:50:29.658642+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.11672v3","created_at":"2026-07-05T10:50:29.658642+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.11672","created_at":"2026-07-05T10:50:29.658642+00:00"},{"alias_kind":"pith_short_12","alias_value":"2LYOCNG623XR","created_at":"2026-07-05T10:50:29.658642+00:00"},{"alias_kind":"pith_short_16","alias_value":"2LYOCNG623XRAJDQ","created_at":"2026-07-05T10:50:29.658642+00:00"},{"alias_kind":"pith_short_8","alias_value":"2LYOCNG6","created_at":"2026-07-05T10:50:29.658642+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.07707","citing_title":"Co-LMLM: Continuous-Query Limited Memory Language Models","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2606.07711","citing_title":"Rosetta Memory: Adaptive Memory for Cross-LLM Agents","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01224","citing_title":"AutoMem: Automated Learning of Memory as a Cognitive Skill","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2405.13068","citing_title":"Uncovering Logit Suppression Vulnerabilities in LLM Safety Alignment","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2601.11848","citing_title":"Compass vs Railway Tracks: Unpacking User Mental Models for Communicating Long-Horizon Work to Humans vs. AI","ref_index":51,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2LYOCNG623XRAJDQJQAC5NFGIM","json":"https://pith.science/pith/2LYOCNG623XRAJDQJQAC5NFGIM.json","graph_json":"https://pith.science/api/pith-number/2LYOCNG623XRAJDQJQAC5NFGIM/graph.json","events_json":"https://pith.science/api/pith-number/2LYOCNG623XRAJDQJQAC5NFGIM/events.json","paper":"https://pith.science/paper/2LYOCNG6"},"agent_actions":{"view_html":"https://pith.science/pith/2LYOCNG623XRAJDQJQAC5NFGIM","download_json":"https://pith.science/pith/2LYOCNG623XRAJDQJQAC5NFGIM.json","view_paper":"https://pith.science/paper/2LYOCNG6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.11672&json=true","fetch_graph":"https://pith.science/api/pith-number/2LYOCNG623XRAJDQJQAC5NFGIM/graph.json","fetch_events":"https://pith.science/api/pith-number/2LYOCNG623XRAJDQJQAC5NFGIM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2LYOCNG623XRAJDQJQAC5NFGIM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2LYOCNG623XRAJDQJQAC5NFGIM/action/storage_attestation","attest_author":"https://pith.science/pith/2LYOCNG623XRAJDQJQAC5NFGIM/action/author_attestation","sign_citation":"https://pith.science/pith/2LYOCNG623XRAJDQJQAC5NFGIM/action/citation_signature","submit_replication":"https://pith.science/pith/2LYOCNG623XRAJDQJQAC5NFGIM/action/replication_record"}},"created_at":"2026-07-05T10:50:29.658642+00:00","updated_at":"2026-07-05T10:50:29.658642+00:00"}