{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:VSQFCPLSRBPFGKHE6TT2U6CTJE","short_pith_number":"pith:VSQFCPLS","canonical_record":{"source":{"id":"2404.08509","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2024-04-12T14:46:15Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"573bcc6daf20d4f70877feb4cc29dc80846553839de2d0601ff16567b77dae53","abstract_canon_sha256":"1c304349f937499bad435658099bb14c8bb6c96a5f22a0bc79f829e342e76980"},"schema_version":"1.0"},"canonical_sha256":"aca0513d72885e5328e4f4e7aa7853491d6f01cea8d2491d361b3805022c49d5","source":{"kind":"arxiv","id":"2404.08509","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.08509","created_at":"2026-07-05T09:39:54Z"},{"alias_kind":"arxiv_version","alias_value":"2404.08509v2","created_at":"2026-07-05T09:39:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.08509","created_at":"2026-07-05T09:39:54Z"},{"alias_kind":"pith_short_12","alias_value":"VSQFCPLSRBPF","created_at":"2026-07-05T09:39:54Z"},{"alias_kind":"pith_short_16","alias_value":"VSQFCPLSRBPFGKHE","created_at":"2026-07-05T09:39:54Z"},{"alias_kind":"pith_short_8","alias_value":"VSQFCPLS","created_at":"2026-07-05T09:39:54Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:VSQFCPLSRBPFGKHE6TT2U6CTJE","target":"record","payload":{"canonical_record":{"source":{"id":"2404.08509","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2024-04-12T14:46:15Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"573bcc6daf20d4f70877feb4cc29dc80846553839de2d0601ff16567b77dae53","abstract_canon_sha256":"1c304349f937499bad435658099bb14c8bb6c96a5f22a0bc79f829e342e76980"},"schema_version":"1.0"},"canonical_sha256":"aca0513d72885e5328e4f4e7aa7853491d6f01cea8d2491d361b3805022c49d5","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:39:54.180174Z","signature_b64":"dxYu3gn1D4k5qdrSrYxYknNvdDxiQHmv2mE8ijPPtU2/0SBH7xAhBHmgnicaI8M+rPxZegNf06miMyxO4aXPBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aca0513d72885e5328e4f4e7aa7853491d6f01cea8d2491d361b3805022c49d5","last_reissued_at":"2026-07-05T09:39:54.179561Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:39:54.179561Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2404.08509","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:39:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Rgdkpio5HWjuqK8InFWvcAv7J3BWGilOMFVAUpNNYKj/1s+H1c38vtkcyIIvGHtq4Bj4oiDtwhiHCo9pGel4DA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T11:27:45.758830Z"},"content_sha256":"a4fff7ba72234448e95122a57a3bf0483005948663e5cec6cd5a87cd803061f5","schema_version":"1.0","event_id":"sha256:a4fff7ba72234448e95122a57a3bf0483005948663e5cec6cd5a87cd803061f5"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:VSQFCPLSRBPFGKHE6TT2U6CTJE","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Efficient Interactive LLM Serving with Proxy Model-based Sequence Length Prediction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.DC","authors_text":"Archit Patke, Chen Wang, Haoran Qiu, Hubertus Franke, Ravishankar K. Iyer, Saurabh Jha, Shengkun Cui, Tamer Ba\\c{s}ar, Weichao Mao, Zbigniew T. Kalbarczyk","submitted_at":"2024-04-12T14:46:15Z","abstract_excerpt":"Large language models (LLMs) have been driving a new wave of interactive AI applications across numerous domains. However, efficiently serving LLM inference requests is challenging due to their unpredictable execution times originating from the autoregressive nature of generative models. Existing LLM serving systems exploit first-come-first-serve (FCFS) scheduling, suffering from head-of-line blocking issues. To address the non-deterministic nature of LLMs and enable efficient interactive LLM serving, we present a speculative shortest-job-first (SSJF) scheduler that uses a light proxy model to"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.08509","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/2404.08509/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:39:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qCK0//1t/hHMTwHLXXXvVYQlQzeFiy+flznpz1hsEWU2g3gQJYGeDXk2BrwHA4RxF2YY8w86kWpugN+6q2KRAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T11:27:45.759338Z"},"content_sha256":"fcac14624a6476f6d0957cd6b9f2fcd5099215e44a28b5bd5221370cc45888f5","schema_version":"1.0","event_id":"sha256:fcac14624a6476f6d0957cd6b9f2fcd5099215e44a28b5bd5221370cc45888f5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/VSQFCPLSRBPFGKHE6TT2U6CTJE/bundle.json","state_url":"https://pith.science/pith/VSQFCPLSRBPFGKHE6TT2U6CTJE/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/VSQFCPLSRBPFGKHE6TT2U6CTJE/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":"2026-08-06T11:27:45Z","links":{"resolver":"https://pith.science/pith/VSQFCPLSRBPFGKHE6TT2U6CTJE","bundle":"https://pith.science/pith/VSQFCPLSRBPFGKHE6TT2U6CTJE/bundle.json","state":"https://pith.science/pith/VSQFCPLSRBPFGKHE6TT2U6CTJE/state.json","well_known_bundle":"https://pith.science/.well-known/pith/VSQFCPLSRBPFGKHE6TT2U6CTJE/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:VSQFCPLSRBPFGKHE6TT2U6CTJE","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":"1c304349f937499bad435658099bb14c8bb6c96a5f22a0bc79f829e342e76980","cross_cats_sorted":["cs.CL","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2024-04-12T14:46:15Z","title_canon_sha256":"573bcc6daf20d4f70877feb4cc29dc80846553839de2d0601ff16567b77dae53"},"schema_version":"1.0","source":{"id":"2404.08509","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.08509","created_at":"2026-07-05T09:39:54Z"},{"alias_kind":"arxiv_version","alias_value":"2404.08509v2","created_at":"2026-07-05T09:39:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.08509","created_at":"2026-07-05T09:39:54Z"},{"alias_kind":"pith_short_12","alias_value":"VSQFCPLSRBPF","created_at":"2026-07-05T09:39:54Z"},{"alias_kind":"pith_short_16","alias_value":"VSQFCPLSRBPFGKHE","created_at":"2026-07-05T09:39:54Z"},{"alias_kind":"pith_short_8","alias_value":"VSQFCPLS","created_at":"2026-07-05T09:39:54Z"}],"graph_snapshots":[{"event_id":"sha256:fcac14624a6476f6d0957cd6b9f2fcd5099215e44a28b5bd5221370cc45888f5","target":"graph","created_at":"2026-07-05T09:39:54Z","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"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2404.08509/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) have been driving a new wave of interactive AI applications across numerous domains. However, efficiently serving LLM inference requests is challenging due to their unpredictable execution times originating from the autoregressive nature of generative models. Existing LLM serving systems exploit first-come-first-serve (FCFS) scheduling, suffering from head-of-line blocking issues. To address the non-deterministic nature of LLMs and enable efficient interactive LLM serving, we present a speculative shortest-job-first (SSJF) scheduler that uses a light proxy model to","authors_text":"Archit Patke, Chen Wang, Haoran Qiu, Hubertus Franke, Ravishankar K. Iyer, Saurabh Jha, Shengkun Cui, Tamer Ba\\c{s}ar, Weichao Mao, Zbigniew T. Kalbarczyk","cross_cats":["cs.CL","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2024-04-12T14:46:15Z","title":"Efficient Interactive LLM Serving with Proxy Model-based Sequence Length Prediction"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.08509","kind":"arxiv","version":2},"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:a4fff7ba72234448e95122a57a3bf0483005948663e5cec6cd5a87cd803061f5","target":"record","created_at":"2026-07-05T09:39:54Z","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":"1c304349f937499bad435658099bb14c8bb6c96a5f22a0bc79f829e342e76980","cross_cats_sorted":["cs.CL","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2024-04-12T14:46:15Z","title_canon_sha256":"573bcc6daf20d4f70877feb4cc29dc80846553839de2d0601ff16567b77dae53"},"schema_version":"1.0","source":{"id":"2404.08509","kind":"arxiv","version":2}},"canonical_sha256":"aca0513d72885e5328e4f4e7aa7853491d6f01cea8d2491d361b3805022c49d5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"aca0513d72885e5328e4f4e7aa7853491d6f01cea8d2491d361b3805022c49d5","first_computed_at":"2026-07-05T09:39:54.179561Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:39:54.179561Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"dxYu3gn1D4k5qdrSrYxYknNvdDxiQHmv2mE8ijPPtU2/0SBH7xAhBHmgnicaI8M+rPxZegNf06miMyxO4aXPBw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:39:54.180174Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.08509","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a4fff7ba72234448e95122a57a3bf0483005948663e5cec6cd5a87cd803061f5","sha256:fcac14624a6476f6d0957cd6b9f2fcd5099215e44a28b5bd5221370cc45888f5"],"state_sha256":"1575ca093e14ec80b100bea1cb9c02f23fc13b112ac792a54d3d7a16f95c65aa"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"LvaHsWHVBjLOFpoZkmnNTvqqWkmnEYdlNu11uJxNN/NhxCokbuZ/7ZtCHjWoCnSVEl4x1Nj57dv8EOI2nUgqCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T11:27:45.762987Z","bundle_sha256":"6ccf24f7447ffc942d3007e46c1fe004517c9bd0d6bd911e461b6a57b06418ad"}}