{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BIJIKZ32EKXZMN32CLFHWRKW2B","short_pith_number":"pith:BIJIKZ32","schema_version":"1.0","canonical_sha256":"0a1285677a22af96377a12ca7b4556d06c8d4d33bbb0fc301083eb67d77e0f10","source":{"kind":"arxiv","id":"2509.02377","version":1},"attestation_state":"computed","paper":{"title":"Upcycling Candidate Tokens of Large Language Models for Query Expansion","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.IR","authors_text":"Jinseok Kim, Sangyeop Kim, Soyeong Jeong, Sukmin Cho, Sungzoon Cho","submitted_at":"2025-09-02T14:43:59Z","abstract_excerpt":"Query Expansion (QE) improves retrieval performance by enriching queries with related terms. Recently, Large Language Models (LLMs) have been used for QE, but existing methods face a trade-off: generating diverse terms boosts performance but increases computational cost. To address this challenge, we propose Candidate Token Query Expansion (CTQE), which extracts diverse and relevant terms from a single LLM decoding pass by leveraging unselected candidate tokens. These tokens, though not part of the final output, are conditioned on the full query and capture useful information. By aggregating t"},"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":"2509.02377","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2025-09-02T14:43:59Z","cross_cats_sorted":[],"title_canon_sha256":"4d6cb6c39b61418c94879727a16e8b946ce4665629a045f59f3155d40f1eede2","abstract_canon_sha256":"614a50aa8d957bf0cbec21527ed90779ffb38e665fd22bf2a06b32514cd1ab53"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:03:37.343138Z","signature_b64":"rlOMFs5uUlesmnU0sa9ANzoRxitC1BpxiC/HXpAhfbZvR37J/pH+x2T1c9ahB5X9v/KNUL0DopLwOKL0DJz5DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0a1285677a22af96377a12ca7b4556d06c8d4d33bbb0fc301083eb67d77e0f10","last_reissued_at":"2026-07-05T12:03:37.342549Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:03:37.342549Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Upcycling Candidate Tokens of Large Language Models for Query Expansion","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.IR","authors_text":"Jinseok Kim, Sangyeop Kim, Soyeong Jeong, Sukmin Cho, Sungzoon Cho","submitted_at":"2025-09-02T14:43:59Z","abstract_excerpt":"Query Expansion (QE) improves retrieval performance by enriching queries with related terms. Recently, Large Language Models (LLMs) have been used for QE, but existing methods face a trade-off: generating diverse terms boosts performance but increases computational cost. To address this challenge, we propose Candidate Token Query Expansion (CTQE), which extracts diverse and relevant terms from a single LLM decoding pass by leveraging unselected candidate tokens. These tokens, though not part of the final output, are conditioned on the full query and capture useful information. By aggregating t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.02377","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/2509.02377/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":"2509.02377","created_at":"2026-07-05T12:03:37.342619+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.02377v1","created_at":"2026-07-05T12:03:37.342619+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.02377","created_at":"2026-07-05T12:03:37.342619+00:00"},{"alias_kind":"pith_short_12","alias_value":"BIJIKZ32EKXZ","created_at":"2026-07-05T12:03:37.342619+00:00"},{"alias_kind":"pith_short_16","alias_value":"BIJIKZ32EKXZMN32","created_at":"2026-07-05T12:03:37.342619+00:00"},{"alias_kind":"pith_short_8","alias_value":"BIJIKZ32","created_at":"2026-07-05T12:03:37.342619+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BIJIKZ32EKXZMN32CLFHWRKW2B","json":"https://pith.science/pith/BIJIKZ32EKXZMN32CLFHWRKW2B.json","graph_json":"https://pith.science/api/pith-number/BIJIKZ32EKXZMN32CLFHWRKW2B/graph.json","events_json":"https://pith.science/api/pith-number/BIJIKZ32EKXZMN32CLFHWRKW2B/events.json","paper":"https://pith.science/paper/BIJIKZ32"},"agent_actions":{"view_html":"https://pith.science/pith/BIJIKZ32EKXZMN32CLFHWRKW2B","download_json":"https://pith.science/pith/BIJIKZ32EKXZMN32CLFHWRKW2B.json","view_paper":"https://pith.science/paper/BIJIKZ32","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.02377&json=true","fetch_graph":"https://pith.science/api/pith-number/BIJIKZ32EKXZMN32CLFHWRKW2B/graph.json","fetch_events":"https://pith.science/api/pith-number/BIJIKZ32EKXZMN32CLFHWRKW2B/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BIJIKZ32EKXZMN32CLFHWRKW2B/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BIJIKZ32EKXZMN32CLFHWRKW2B/action/storage_attestation","attest_author":"https://pith.science/pith/BIJIKZ32EKXZMN32CLFHWRKW2B/action/author_attestation","sign_citation":"https://pith.science/pith/BIJIKZ32EKXZMN32CLFHWRKW2B/action/citation_signature","submit_replication":"https://pith.science/pith/BIJIKZ32EKXZMN32CLFHWRKW2B/action/replication_record"}},"created_at":"2026-07-05T12:03:37.342619+00:00","updated_at":"2026-07-05T12:03:37.342619+00:00"}