{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:67YMS7LAGCSBXGBWZUODUG263B","short_pith_number":"pith:67YMS7LA","schema_version":"1.0","canonical_sha256":"f7f0c97d6030a41b9836cd1c3a1b5ed8734c0e5d319dfb8785c60fd5c08bba51","source":{"kind":"arxiv","id":"2312.05516","version":3},"attestation_state":"computed","paper":{"title":"Stateful Large Language Model Serving with Pensieve","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DC"],"primary_cat":"cs.LG","authors_text":"Jinkun Lin, Jinyang Li, Lingfan Yu","submitted_at":"2023-12-09T09:55:07Z","abstract_excerpt":"Large Language Models (LLMs) are wildly popular today and it is important to serve them efficiently. Existing LLM serving systems are stateless across requests. Consequently, when LLMs are used in the common setting of multi-turn conversations, a growing log of the conversation history must be processed alongside any request by the serving system at each turn, resulting in repeated processing.\n  In this paper, we design $Pensieve$, a system optimized for multi-turn conversation LLM serving. $Pensieve$ maintains the conversation state across requests by caching previously processed history to a"},"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":"2312.05516","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-12-09T09:55:07Z","cross_cats_sorted":["cs.DC"],"title_canon_sha256":"29661d1ed78065cf06d5796e4979dd43bfd4784208cef8f4d159951c9fa7fd0d","abstract_canon_sha256":"01a5aa2a4078c35c2ffd1a91df9fd5254a4f63644c233b0b3009a4121ecfd1c8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:16:45.709099Z","signature_b64":"bY5hwUZAQfouLXRmZxRT7Im9uRmY4HfPd5hH1d/F0dIu6iy15bIsrCYZ4DgTZ7cSfZPHFf3dp6dkTvcQOrieDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f7f0c97d6030a41b9836cd1c3a1b5ed8734c0e5d319dfb8785c60fd5c08bba51","last_reissued_at":"2026-07-05T09:16:45.708517Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:16:45.708517Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Stateful Large Language Model Serving with Pensieve","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DC"],"primary_cat":"cs.LG","authors_text":"Jinkun Lin, Jinyang Li, Lingfan Yu","submitted_at":"2023-12-09T09:55:07Z","abstract_excerpt":"Large Language Models (LLMs) are wildly popular today and it is important to serve them efficiently. Existing LLM serving systems are stateless across requests. Consequently, when LLMs are used in the common setting of multi-turn conversations, a growing log of the conversation history must be processed alongside any request by the serving system at each turn, resulting in repeated processing.\n  In this paper, we design $Pensieve$, a system optimized for multi-turn conversation LLM serving. $Pensieve$ maintains the conversation state across requests by caching previously processed history to a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.05516","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/2312.05516/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":"2312.05516","created_at":"2026-07-05T09:16:45.708601+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.05516v3","created_at":"2026-07-05T09:16:45.708601+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.05516","created_at":"2026-07-05T09:16:45.708601+00:00"},{"alias_kind":"pith_short_12","alias_value":"67YMS7LAGCSB","created_at":"2026-07-05T09:16:45.708601+00:00"},{"alias_kind":"pith_short_16","alias_value":"67YMS7LAGCSBXGBW","created_at":"2026-07-05T09:16:45.708601+00:00"},{"alias_kind":"pith_short_8","alias_value":"67YMS7LA","created_at":"2026-07-05T09:16:45.708601+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08565","citing_title":"SMetric: Rethink LLM Scheduling for Serving Agents with Balanced Session-centric Scheduling","ref_index":53,"is_internal_anchor":true},{"citing_arxiv_id":"2606.20537","citing_title":"Execution-State Capsules: Graph-Bound Execution-State Checkpoint and Restore for Low-Latency, Small-Batch, On-Device Physical-AI Serving","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12556","citing_title":"ITME: Inference Tiered Memory Expansion with Disaggregated CXL-Hybrid Memories","ref_index":43,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/67YMS7LAGCSBXGBWZUODUG263B","json":"https://pith.science/pith/67YMS7LAGCSBXGBWZUODUG263B.json","graph_json":"https://pith.science/api/pith-number/67YMS7LAGCSBXGBWZUODUG263B/graph.json","events_json":"https://pith.science/api/pith-number/67YMS7LAGCSBXGBWZUODUG263B/events.json","paper":"https://pith.science/paper/67YMS7LA"},"agent_actions":{"view_html":"https://pith.science/pith/67YMS7LAGCSBXGBWZUODUG263B","download_json":"https://pith.science/pith/67YMS7LAGCSBXGBWZUODUG263B.json","view_paper":"https://pith.science/paper/67YMS7LA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.05516&json=true","fetch_graph":"https://pith.science/api/pith-number/67YMS7LAGCSBXGBWZUODUG263B/graph.json","fetch_events":"https://pith.science/api/pith-number/67YMS7LAGCSBXGBWZUODUG263B/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/67YMS7LAGCSBXGBWZUODUG263B/action/timestamp_anchor","attest_storage":"https://pith.science/pith/67YMS7LAGCSBXGBWZUODUG263B/action/storage_attestation","attest_author":"https://pith.science/pith/67YMS7LAGCSBXGBWZUODUG263B/action/author_attestation","sign_citation":"https://pith.science/pith/67YMS7LAGCSBXGBWZUODUG263B/action/citation_signature","submit_replication":"https://pith.science/pith/67YMS7LAGCSBXGBWZUODUG263B/action/replication_record"}},"created_at":"2026-07-05T09:16:45.708601+00:00","updated_at":"2026-07-05T09:16:45.708601+00:00"}