{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:QR7S6ONBOBXJKQZ57LR2OYGHDV","short_pith_number":"pith:QR7S6ONB","schema_version":"1.0","canonical_sha256":"847f2f39a1706e95433dfae3a760c71d67ec8b71a3a7a53053f3e54d024aa7ad","source":{"kind":"arxiv","id":"2406.08477","version":1},"attestation_state":"computed","paper":{"title":"Improving LLMs for Recommendation with Out-Of-Vocabulary Tokens","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.IR","authors_text":"Chunxu Shen, De-Chuan Zhan, Han-Jia Ye, Jia-Qi Yang, Kai-Qi Liu, Ting-Ji Huang","submitted_at":"2024-06-12T17:59:05Z","abstract_excerpt":"Characterizing users and items through vector representations is crucial for various tasks in recommender systems. Recent approaches attempt to apply Large Language Models (LLMs) in recommendation through a question and answer format, where real users and items (e.g., Item No.2024) are represented with in-vocabulary tokens (e.g., \"item\", \"20\", \"24\"). However, since LLMs are typically pretrained on natural language tasks, these in-vocabulary tokens lack the expressive power for distinctive users and items, thereby weakening the recommendation ability even after fine-tuning on recommendation tas"},"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":"2406.08477","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.IR","submitted_at":"2024-06-12T17:59:05Z","cross_cats_sorted":[],"title_canon_sha256":"5b59a76fcf3bc4b3b0e547b96012d27268ba63415b95588ded0dd6c87eef4472","abstract_canon_sha256":"db98892b47c9f256c952c426ee6316779a87402788cb82677d3ee69839001b4f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:31:00.553879Z","signature_b64":"zUw20Al3ypB5SKV/WZXZnSD7EwHyMpGLo0FeQ8IDdaANSiqj9k7tpP8TDuNVFLwG6s+z7NzXb34JtB/1p7sJBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"847f2f39a1706e95433dfae3a760c71d67ec8b71a3a7a53053f3e54d024aa7ad","last_reissued_at":"2026-07-05T08:31:00.553259Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:31:00.553259Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving LLMs for Recommendation with Out-Of-Vocabulary Tokens","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.IR","authors_text":"Chunxu Shen, De-Chuan Zhan, Han-Jia Ye, Jia-Qi Yang, Kai-Qi Liu, Ting-Ji Huang","submitted_at":"2024-06-12T17:59:05Z","abstract_excerpt":"Characterizing users and items through vector representations is crucial for various tasks in recommender systems. Recent approaches attempt to apply Large Language Models (LLMs) in recommendation through a question and answer format, where real users and items (e.g., Item No.2024) are represented with in-vocabulary tokens (e.g., \"item\", \"20\", \"24\"). However, since LLMs are typically pretrained on natural language tasks, these in-vocabulary tokens lack the expressive power for distinctive users and items, thereby weakening the recommendation ability even after fine-tuning on recommendation tas"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.08477","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/2406.08477/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":"2406.08477","created_at":"2026-07-05T08:31:00.553343+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.08477v1","created_at":"2026-07-05T08:31:00.553343+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.08477","created_at":"2026-07-05T08:31:00.553343+00:00"},{"alias_kind":"pith_short_12","alias_value":"QR7S6ONBOBXJ","created_at":"2026-07-05T08:31:00.553343+00:00"},{"alias_kind":"pith_short_16","alias_value":"QR7S6ONBOBXJKQZ5","created_at":"2026-07-05T08:31:00.553343+00:00"},{"alias_kind":"pith_short_8","alias_value":"QR7S6ONB","created_at":"2026-07-05T08:31:00.553343+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.20490","citing_title":"Break the Optimization Barrier of LLM-Enhanced Recommenders: A Theoretical Analysis and Practical Framework","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QR7S6ONBOBXJKQZ57LR2OYGHDV","json":"https://pith.science/pith/QR7S6ONBOBXJKQZ57LR2OYGHDV.json","graph_json":"https://pith.science/api/pith-number/QR7S6ONBOBXJKQZ57LR2OYGHDV/graph.json","events_json":"https://pith.science/api/pith-number/QR7S6ONBOBXJKQZ57LR2OYGHDV/events.json","paper":"https://pith.science/paper/QR7S6ONB"},"agent_actions":{"view_html":"https://pith.science/pith/QR7S6ONBOBXJKQZ57LR2OYGHDV","download_json":"https://pith.science/pith/QR7S6ONBOBXJKQZ57LR2OYGHDV.json","view_paper":"https://pith.science/paper/QR7S6ONB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.08477&json=true","fetch_graph":"https://pith.science/api/pith-number/QR7S6ONBOBXJKQZ57LR2OYGHDV/graph.json","fetch_events":"https://pith.science/api/pith-number/QR7S6ONBOBXJKQZ57LR2OYGHDV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QR7S6ONBOBXJKQZ57LR2OYGHDV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QR7S6ONBOBXJKQZ57LR2OYGHDV/action/storage_attestation","attest_author":"https://pith.science/pith/QR7S6ONBOBXJKQZ57LR2OYGHDV/action/author_attestation","sign_citation":"https://pith.science/pith/QR7S6ONBOBXJKQZ57LR2OYGHDV/action/citation_signature","submit_replication":"https://pith.science/pith/QR7S6ONBOBXJKQZ57LR2OYGHDV/action/replication_record"}},"created_at":"2026-07-05T08:31:00.553343+00:00","updated_at":"2026-07-05T08:31:00.553343+00:00"}