{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YAWIU6OS7GFGI67WQHA4UUAGE7","short_pith_number":"pith:YAWIU6OS","schema_version":"1.0","canonical_sha256":"c02c8a79d2f98a647bf681c1ca500627c191ed0ded60e8e231375cf9b3d1fc9e","source":{"kind":"arxiv","id":"2502.07776","version":2},"attestation_state":"computed","paper":{"title":"Auditing Prompt Caching in Language Model APIs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chenchen Gu, Percy Liang, Rohith Kuditipudi, Tatsunori Hashimoto, Xiang Lisa Li","submitted_at":"2025-02-11T18:58:04Z","abstract_excerpt":"Prompt caching in large language models (LLMs) results in data-dependent timing variations: cached prompts are processed faster than non-cached prompts. These timing differences introduce the risk of side-channel timing attacks. For example, if the cache is shared across users, an attacker could identify cached prompts from fast API response times to learn information about other users' prompts. Because prompt caching may cause privacy leakage, transparency around the caching policies of API providers is important. To this end, we develop and conduct statistical audits to detect prompt caching"},"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":"2502.07776","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-11T18:58:04Z","cross_cats_sorted":["cs.CR","cs.LG"],"title_canon_sha256":"1605ed9f46f93975bd22cebb13fe21d6071fe56d9fb869600d854c0d0d5f6a6a","abstract_canon_sha256":"45349ff25b7be7b4375165d48d51e898bf3edf78fba00989224c949c48ebfc46"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:35:53.336572Z","signature_b64":"j8yg4qBlxXsRC2rPDhtYx2rYCjSe0ja93O/UZduW5SG/I3JPMrdiqBJyiEBR1OR8UoZekVdoi/RMOfHvbsVZAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c02c8a79d2f98a647bf681c1ca500627c191ed0ded60e8e231375cf9b3d1fc9e","last_reissued_at":"2026-07-05T11:35:53.336068Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:35:53.336068Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Auditing Prompt Caching in Language Model APIs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chenchen Gu, Percy Liang, Rohith Kuditipudi, Tatsunori Hashimoto, Xiang Lisa Li","submitted_at":"2025-02-11T18:58:04Z","abstract_excerpt":"Prompt caching in large language models (LLMs) results in data-dependent timing variations: cached prompts are processed faster than non-cached prompts. These timing differences introduce the risk of side-channel timing attacks. For example, if the cache is shared across users, an attacker could identify cached prompts from fast API response times to learn information about other users' prompts. Because prompt caching may cause privacy leakage, transparency around the caching policies of API providers is important. To this end, we develop and conduct statistical audits to detect prompt caching"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.07776","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/2502.07776/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":"2502.07776","created_at":"2026-07-05T11:35:53.336127+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.07776v2","created_at":"2026-07-05T11:35:53.336127+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.07776","created_at":"2026-07-05T11:35:53.336127+00:00"},{"alias_kind":"pith_short_12","alias_value":"YAWIU6OS7GFG","created_at":"2026-07-05T11:35:53.336127+00:00"},{"alias_kind":"pith_short_16","alias_value":"YAWIU6OS7GFGI67W","created_at":"2026-07-05T11:35:53.336127+00:00"},{"alias_kind":"pith_short_8","alias_value":"YAWIU6OS","created_at":"2026-07-05T11:35:53.336127+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.11470","citing_title":"The Periodic Table of LLM Reasoning: A Structured Survey of Reasoning Paradigms, Methods, and Failure Modes","ref_index":80,"is_internal_anchor":false},{"citing_arxiv_id":"2503.17181","citing_title":"A Study of LLMs' Preferences for Libraries and Programming Languages","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2603.10726","citing_title":"PrefixWall: Mitigating Prefix Caching Side Channels in Shared LLM Systems","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2603.09002","citing_title":"Security Considerations for Multi-agent Systems","ref_index":224,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21083","citing_title":"Behavioral Consistency and Transparency Analysis on Large Language Model API Gateways","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YAWIU6OS7GFGI67WQHA4UUAGE7","json":"https://pith.science/pith/YAWIU6OS7GFGI67WQHA4UUAGE7.json","graph_json":"https://pith.science/api/pith-number/YAWIU6OS7GFGI67WQHA4UUAGE7/graph.json","events_json":"https://pith.science/api/pith-number/YAWIU6OS7GFGI67WQHA4UUAGE7/events.json","paper":"https://pith.science/paper/YAWIU6OS"},"agent_actions":{"view_html":"https://pith.science/pith/YAWIU6OS7GFGI67WQHA4UUAGE7","download_json":"https://pith.science/pith/YAWIU6OS7GFGI67WQHA4UUAGE7.json","view_paper":"https://pith.science/paper/YAWIU6OS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.07776&json=true","fetch_graph":"https://pith.science/api/pith-number/YAWIU6OS7GFGI67WQHA4UUAGE7/graph.json","fetch_events":"https://pith.science/api/pith-number/YAWIU6OS7GFGI67WQHA4UUAGE7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YAWIU6OS7GFGI67WQHA4UUAGE7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YAWIU6OS7GFGI67WQHA4UUAGE7/action/storage_attestation","attest_author":"https://pith.science/pith/YAWIU6OS7GFGI67WQHA4UUAGE7/action/author_attestation","sign_citation":"https://pith.science/pith/YAWIU6OS7GFGI67WQHA4UUAGE7/action/citation_signature","submit_replication":"https://pith.science/pith/YAWIU6OS7GFGI67WQHA4UUAGE7/action/replication_record"}},"created_at":"2026-07-05T11:35:53.336127+00:00","updated_at":"2026-07-05T11:35:53.336127+00:00"}