{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:YMQXB3EEEKOIJP25NJMBWMIDNT","short_pith_number":"pith:YMQXB3EE","canonical_record":{"source":{"id":"2605.15051","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T16:45:28Z","cross_cats_sorted":["cs.PF"],"title_canon_sha256":"09ce039476b327e6af760e82455096599ced97386b21fd7353289c4258cd2ae6","abstract_canon_sha256":"9ce0c07852db24d1e66d94d071e21022148fe8a670c96e165d2cb8c46ddc50c6"},"schema_version":"1.0"},"canonical_sha256":"c32170ec84229c84bf5d6a581b31036cc2f9fef984644800bd7286d2f3cd151c","source":{"kind":"arxiv","id":"2605.15051","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15051","created_at":"2026-05-17T23:38:54Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15051v1","created_at":"2026-05-17T23:38:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15051","created_at":"2026-05-17T23:38:54Z"},{"alias_kind":"pith_short_12","alias_value":"YMQXB3EEEKOI","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"YMQXB3EEEKOIJP25","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"YMQXB3EE","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:YMQXB3EEEKOIJP25NJMBWMIDNT","target":"record","payload":{"canonical_record":{"source":{"id":"2605.15051","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T16:45:28Z","cross_cats_sorted":["cs.PF"],"title_canon_sha256":"09ce039476b327e6af760e82455096599ced97386b21fd7353289c4258cd2ae6","abstract_canon_sha256":"9ce0c07852db24d1e66d94d071e21022148fe8a670c96e165d2cb8c46ddc50c6"},"schema_version":"1.0"},"canonical_sha256":"c32170ec84229c84bf5d6a581b31036cc2f9fef984644800bd7286d2f3cd151c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:54.388129Z","signature_b64":"7AdIiJ73L+uH57gge4NUJu22UigOh7z11KGpaIbtKbtfv7ZaPKGN0LcfN82NFE7aZhuorDBWCaBH3sCO69nMBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c32170ec84229c84bf5d6a581b31036cc2f9fef984644800bd7286d2f3cd151c","last_reissued_at":"2026-05-17T23:38:54.387416Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:54.387416Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.15051","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"soqhJylsx8dKxlXTzZWEwlFWG65psdoc71Vf6RCTBnYp69d4NteFxKhflSdOGHKuHhnopfXjiCyhJdsL0j0/Dg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-30T07:17:44.188836Z"},"content_sha256":"fdbacaf5dcc341f8dcb76e446c854cae19697018cf3176acec6f6682252a5c75","schema_version":"1.0","event_id":"sha256:fdbacaf5dcc341f8dcb76e446c854cae19697018cf3176acec6f6682252a5c75"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:YMQXB3EEEKOIJP25NJMBWMIDNT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"An Interpretable Latency Model for Speculative Decoding in LLM Serving","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.PF"],"primary_cat":"cs.LG","authors_text":"Alexandre Marques, Linghao Kong, Mark Kurtz, Megan Flynn, Michael Peng, Nir Shavit","submitted_at":"2026-05-14T16:45:28Z","abstract_excerpt":"Speculative decoding (SD) accelerates large language model (LLM) inference by using a smaller draft model to propose multiple tokens that are verified by a larger target model in parallel. While prior work demonstrates substantial speedups in isolated or fixed-batch settings, the behavior of SD in production serving systems remains poorly understood: request load varies over time, and effective batch size emerges from the serving system rather than being directly controlled or observed. In this work, we develop a simple and interpretable latency model for SD in LLM serving. We infer effective "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.15051","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":""},"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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Cc/TA9MNCFMX/zJz1/cFTyjQT9nV+vP+iHDICLpTig0ZZjHBHtAoFUKYF66e4OPOjzMGxPAQNsLdn9HHmSFdAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-30T07:17:44.189193Z"},"content_sha256":"f81c20e7fbe738e02b139c693293924fe92a6529a19a03466e53690f3b45d765","schema_version":"1.0","event_id":"sha256:f81c20e7fbe738e02b139c693293924fe92a6529a19a03466e53690f3b45d765"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/YMQXB3EEEKOIJP25NJMBWMIDNT/bundle.json","state_url":"https://pith.science/pith/YMQXB3EEEKOIJP25NJMBWMIDNT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/YMQXB3EEEKOIJP25NJMBWMIDNT/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-30T07:17:44Z","links":{"resolver":"https://pith.science/pith/YMQXB3EEEKOIJP25NJMBWMIDNT","bundle":"https://pith.science/pith/YMQXB3EEEKOIJP25NJMBWMIDNT/bundle.json","state":"https://pith.science/pith/YMQXB3EEEKOIJP25NJMBWMIDNT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/YMQXB3EEEKOIJP25NJMBWMIDNT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:YMQXB3EEEKOIJP25NJMBWMIDNT","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"9ce0c07852db24d1e66d94d071e21022148fe8a670c96e165d2cb8c46ddc50c6","cross_cats_sorted":["cs.PF"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T16:45:28Z","title_canon_sha256":"09ce039476b327e6af760e82455096599ced97386b21fd7353289c4258cd2ae6"},"schema_version":"1.0","source":{"id":"2605.15051","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15051","created_at":"2026-05-17T23:38:54Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15051v1","created_at":"2026-05-17T23:38:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15051","created_at":"2026-05-17T23:38:54Z"},{"alias_kind":"pith_short_12","alias_value":"YMQXB3EEEKOI","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"YMQXB3EEEKOIJP25","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"YMQXB3EE","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:f81c20e7fbe738e02b139c693293924fe92a6529a19a03466e53690f3b45d765","target":"graph","created_at":"2026-05-17T23:38:54Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Speculative decoding (SD) accelerates large language model (LLM) inference by using a smaller draft model to propose multiple tokens that are verified by a larger target model in parallel. While prior work demonstrates substantial speedups in isolated or fixed-batch settings, the behavior of SD in production serving systems remains poorly understood: request load varies over time, and effective batch size emerges from the serving system rather than being directly controlled or observed. In this work, we develop a simple and interpretable latency model for SD in LLM serving. We infer effective ","authors_text":"Alexandre Marques, Linghao Kong, Mark Kurtz, Megan Flynn, Michael Peng, Nir Shavit","cross_cats":["cs.PF"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T16:45:28Z","title":"An Interpretable Latency Model for Speculative Decoding in LLM Serving"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.15051","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:fdbacaf5dcc341f8dcb76e446c854cae19697018cf3176acec6f6682252a5c75","target":"record","created_at":"2026-05-17T23:38:54Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"9ce0c07852db24d1e66d94d071e21022148fe8a670c96e165d2cb8c46ddc50c6","cross_cats_sorted":["cs.PF"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T16:45:28Z","title_canon_sha256":"09ce039476b327e6af760e82455096599ced97386b21fd7353289c4258cd2ae6"},"schema_version":"1.0","source":{"id":"2605.15051","kind":"arxiv","version":1}},"canonical_sha256":"c32170ec84229c84bf5d6a581b31036cc2f9fef984644800bd7286d2f3cd151c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c32170ec84229c84bf5d6a581b31036cc2f9fef984644800bd7286d2f3cd151c","first_computed_at":"2026-05-17T23:38:54.387416Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:54.387416Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"7AdIiJ73L+uH57gge4NUJu22UigOh7z11KGpaIbtKbtfv7ZaPKGN0LcfN82NFE7aZhuorDBWCaBH3sCO69nMBg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:54.388129Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.15051","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:fdbacaf5dcc341f8dcb76e446c854cae19697018cf3176acec6f6682252a5c75","sha256:f81c20e7fbe738e02b139c693293924fe92a6529a19a03466e53690f3b45d765"],"state_sha256":"54270d3bd5c1db54eb13c304f61005378bc5f6d1aeb4a0294c939cab9998c4f1"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tlU5HTy/9URQ/hDhNsEiZeiYpMS52xeuu676DEexBnMgUNwcPc0nMxtMhkaTdhOrA8Uw6MCav8ORyV1fWdX4Aw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-30T07:17:44.191269Z","bundle_sha256":"7bea3b3c3598150e0bf122a423df142850aff1462d8dc8ef580fe8c7a9abd058"}}