{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:UVDFX2UO2JHY6OQW5S37ZK7GMC","short_pith_number":"pith:UVDFX2UO","schema_version":"1.0","canonical_sha256":"a5465bea8ed24f8f3a16ecb7fcabe660a56b1a9fb1161aa9c54cd76eb30dbe8a","source":{"kind":"arxiv","id":"2410.08014","version":2},"attestation_state":"computed","paper":{"title":"Privacy-preserved LLM Cascade via CoT-enhanced Policy Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alec Go, Congchao Wang, Kai Zhang, Liqian Peng, Xiaozhong Liu","submitted_at":"2024-10-10T15:09:52Z","abstract_excerpt":"Large Language Models (LLMs) have gained significant attention in on-device applications due to their remarkable performance across real-world tasks. However, on-device LLMs often suffer from suboptimal performance due to hardware limitations. A promising solution to this challenge is cascading a weaker local (on-device) LLM with a more powerful server LLM. While existing research on LLM cascade primarily optimizes the performance-cost trade-off, real-world applications impose additional requirements, such as privacy preservation, which remain largely unaddressed. In this work, we move beyond "},"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":"2410.08014","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-10T15:09:52Z","cross_cats_sorted":[],"title_canon_sha256":"b1cf1e5268619eac90fcd4c6260169bc6eab2760a1f6991a5853d799283c65fa","abstract_canon_sha256":"872f178dd426681bafd00ebbd34d496ba514c27257694698f1328c8e8da2140d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:21:24.386170Z","signature_b64":"BzzwyAG7qZ5wMQdxfdvu5/9/XMef3eWgDr7TDSgO/1xJ0liFQWo6YO+mGId5hFhg1pYfIamfmnFL3Da6mF+9Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a5465bea8ed24f8f3a16ecb7fcabe660a56b1a9fb1161aa9c54cd76eb30dbe8a","last_reissued_at":"2026-07-05T10:21:24.385510Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:21:24.385510Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Privacy-preserved LLM Cascade via CoT-enhanced Policy Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alec Go, Congchao Wang, Kai Zhang, Liqian Peng, Xiaozhong Liu","submitted_at":"2024-10-10T15:09:52Z","abstract_excerpt":"Large Language Models (LLMs) have gained significant attention in on-device applications due to their remarkable performance across real-world tasks. However, on-device LLMs often suffer from suboptimal performance due to hardware limitations. A promising solution to this challenge is cascading a weaker local (on-device) LLM with a more powerful server LLM. While existing research on LLM cascade primarily optimizes the performance-cost trade-off, real-world applications impose additional requirements, such as privacy preservation, which remain largely unaddressed. In this work, we move beyond "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.08014","kind":"arxiv","version":2},"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/2410.08014/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":"2410.08014","created_at":"2026-07-05T10:21:24.385594+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.08014v2","created_at":"2026-07-05T10:21:24.385594+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.08014","created_at":"2026-07-05T10:21:24.385594+00:00"},{"alias_kind":"pith_short_12","alias_value":"UVDFX2UO2JHY","created_at":"2026-07-05T10:21:24.385594+00:00"},{"alias_kind":"pith_short_16","alias_value":"UVDFX2UO2JHY6OQW","created_at":"2026-07-05T10:21:24.385594+00:00"},{"alias_kind":"pith_short_8","alias_value":"UVDFX2UO","created_at":"2026-07-05T10:21:24.385594+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.03834","citing_title":"Economic Evaluation of LLMs","ref_index":58,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UVDFX2UO2JHY6OQW5S37ZK7GMC","json":"https://pith.science/pith/UVDFX2UO2JHY6OQW5S37ZK7GMC.json","graph_json":"https://pith.science/api/pith-number/UVDFX2UO2JHY6OQW5S37ZK7GMC/graph.json","events_json":"https://pith.science/api/pith-number/UVDFX2UO2JHY6OQW5S37ZK7GMC/events.json","paper":"https://pith.science/paper/UVDFX2UO"},"agent_actions":{"view_html":"https://pith.science/pith/UVDFX2UO2JHY6OQW5S37ZK7GMC","download_json":"https://pith.science/pith/UVDFX2UO2JHY6OQW5S37ZK7GMC.json","view_paper":"https://pith.science/paper/UVDFX2UO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.08014&json=true","fetch_graph":"https://pith.science/api/pith-number/UVDFX2UO2JHY6OQW5S37ZK7GMC/graph.json","fetch_events":"https://pith.science/api/pith-number/UVDFX2UO2JHY6OQW5S37ZK7GMC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UVDFX2UO2JHY6OQW5S37ZK7GMC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UVDFX2UO2JHY6OQW5S37ZK7GMC/action/storage_attestation","attest_author":"https://pith.science/pith/UVDFX2UO2JHY6OQW5S37ZK7GMC/action/author_attestation","sign_citation":"https://pith.science/pith/UVDFX2UO2JHY6OQW5S37ZK7GMC/action/citation_signature","submit_replication":"https://pith.science/pith/UVDFX2UO2JHY6OQW5S37ZK7GMC/action/replication_record"}},"created_at":"2026-07-05T10:21:24.385594+00:00","updated_at":"2026-07-05T10:21:24.385594+00:00"}