{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:VDCYKN25QYIISQTBRX3TVKKUZN","short_pith_number":"pith:VDCYKN25","canonical_record":{"source":{"id":"2607.25852","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-28T15:25:05Z","cross_cats_sorted":[],"title_canon_sha256":"3b2355ddcdc1ece1d22de78ea9bb1b1d766856c44ae4704e5ffa463bd093c6f6","abstract_canon_sha256":"aef2a0a625ec79777aa5a04917b2ba03f969ed1e599bbb3732722c22c666bbd9"},"schema_version":"1.0"},"canonical_sha256":"a8c585375d86108942618df73aa954cb42406a8cec42e0616dfb5da0836bf5a8","source":{"kind":"arxiv","id":"2607.25852","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.25852","created_at":"2026-07-29T01:26:06Z"},{"alias_kind":"arxiv_version","alias_value":"2607.25852v1","created_at":"2026-07-29T01:26:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.25852","created_at":"2026-07-29T01:26:06Z"},{"alias_kind":"pith_short_12","alias_value":"VDCYKN25QYII","created_at":"2026-07-29T01:26:06Z"},{"alias_kind":"pith_short_16","alias_value":"VDCYKN25QYIISQTB","created_at":"2026-07-29T01:26:06Z"},{"alias_kind":"pith_short_8","alias_value":"VDCYKN25","created_at":"2026-07-29T01:26:06Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:VDCYKN25QYIISQTBRX3TVKKUZN","target":"record","payload":{"canonical_record":{"source":{"id":"2607.25852","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-28T15:25:05Z","cross_cats_sorted":[],"title_canon_sha256":"3b2355ddcdc1ece1d22de78ea9bb1b1d766856c44ae4704e5ffa463bd093c6f6","abstract_canon_sha256":"aef2a0a625ec79777aa5a04917b2ba03f969ed1e599bbb3732722c22c666bbd9"},"schema_version":"1.0"},"canonical_sha256":"a8c585375d86108942618df73aa954cb42406a8cec42e0616dfb5da0836bf5a8","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-29T01:26:06.713670Z","signature_b64":"LklIFNUkCppVEJ3SaTUyZyNr0TPtGlnI6FbE7bxpix6nfTfEGqlGzD/Oo9xu3/SZUWWPE7Y4QL5xQwRT/7DHDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a8c585375d86108942618df73aa954cb42406a8cec42e0616dfb5da0836bf5a8","last_reissued_at":"2026-07-29T01:26:06.712832Z","signature_status":"signed_v1","first_computed_at":"2026-07-29T01:26:06.712832Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2607.25852","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":"2026-07-29T01:26:06Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"aUPPPRFX3IBsttUFUA+Rl9Iww7WakIidHTgLmy1AoCvqVX8dT3sPLOWTzVQMG3CbFwVHpuhEkheu0pk1eUl+BQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T08:27:47.615607Z"},"content_sha256":"8f085f5ba2e600b6f50c5cde40f44a81c81cbe4534da02638f770425f3730c40","schema_version":"1.0","event_id":"sha256:8f085f5ba2e600b6f50c5cde40f44a81c81cbe4534da02638f770425f3730c40"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:VDCYKN25QYIISQTBRX3TVKKUZN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AngelSpec: Towards Real-World High Performance Inference with Speculative Decoding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Guanghua Yu, Guangshuo Qin, Guoliang Zhao, Hong Liu, Jianchen Zhu, Jiebin Zhang, Junhan Shi, Kai Zhang, Rui Cen, Runzhi Fan, Ruobing Xie, Song Liu, Tianyu Liu","submitted_at":"2026-07-28T15:25:05Z","abstract_excerpt":"Speculative decoding accelerates large language model inference without changing the target distribution, but no single drafting structure performs best across real-world workloads. Autoregressive multi-token prediction (MTP) is a lightweight, stable proposal mechanism, whereas block-parallel diffusion amortizes drafting latency over much longer candidate sequences; the better choice depends strongly on the output distribution. We present AngelSpec, a unified training framework for MTP and block-parallel speculative decoding that addresses this heterogeneity at three levels. At the training le"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.25852","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2607.25852/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":"2026-07-29T01:26:06Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"UYSmx4ivsJqUVef5LjTctOMCSoO742Q4f+hrUepyZYhzYsZS7cspt55d1IEvMGIqTzRhoYuCTIcxxihft9xADw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T08:27:47.615997Z"},"content_sha256":"134f3a4ae2114b75e9ccda163e4d9d7dc304fc967b4894f11ed3afca7ec8a049","schema_version":"1.0","event_id":"sha256:134f3a4ae2114b75e9ccda163e4d9d7dc304fc967b4894f11ed3afca7ec8a049"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/VDCYKN25QYIISQTBRX3TVKKUZN/bundle.json","state_url":"https://pith.science/pith/VDCYKN25QYIISQTBRX3TVKKUZN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/VDCYKN25QYIISQTBRX3TVKKUZN/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":"2026-08-21T08:27:47Z","links":{"resolver":"https://pith.science/pith/VDCYKN25QYIISQTBRX3TVKKUZN","bundle":"https://pith.science/pith/VDCYKN25QYIISQTBRX3TVKKUZN/bundle.json","state":"https://pith.science/pith/VDCYKN25QYIISQTBRX3TVKKUZN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/VDCYKN25QYIISQTBRX3TVKKUZN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:VDCYKN25QYIISQTBRX3TVKKUZN","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":"aef2a0a625ec79777aa5a04917b2ba03f969ed1e599bbb3732722c22c666bbd9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-28T15:25:05Z","title_canon_sha256":"3b2355ddcdc1ece1d22de78ea9bb1b1d766856c44ae4704e5ffa463bd093c6f6"},"schema_version":"1.0","source":{"id":"2607.25852","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.25852","created_at":"2026-07-29T01:26:06Z"},{"alias_kind":"arxiv_version","alias_value":"2607.25852v1","created_at":"2026-07-29T01:26:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.25852","created_at":"2026-07-29T01:26:06Z"},{"alias_kind":"pith_short_12","alias_value":"VDCYKN25QYII","created_at":"2026-07-29T01:26:06Z"},{"alias_kind":"pith_short_16","alias_value":"VDCYKN25QYIISQTB","created_at":"2026-07-29T01:26:06Z"},{"alias_kind":"pith_short_8","alias_value":"VDCYKN25","created_at":"2026-07-29T01:26:06Z"}],"graph_snapshots":[{"event_id":"sha256:134f3a4ae2114b75e9ccda163e4d9d7dc304fc967b4894f11ed3afca7ec8a049","target":"graph","created_at":"2026-07-29T01:26:06Z","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/2607.25852/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Speculative decoding accelerates large language model inference without changing the target distribution, but no single drafting structure performs best across real-world workloads. Autoregressive multi-token prediction (MTP) is a lightweight, stable proposal mechanism, whereas block-parallel diffusion amortizes drafting latency over much longer candidate sequences; the better choice depends strongly on the output distribution. We present AngelSpec, a unified training framework for MTP and block-parallel speculative decoding that addresses this heterogeneity at three levels. At the training le","authors_text":"Guanghua Yu, Guangshuo Qin, Guoliang Zhao, Hong Liu, Jianchen Zhu, Jiebin Zhang, Junhan Shi, Kai Zhang, Rui Cen, Runzhi Fan, Ruobing Xie, Song Liu, Tianyu Liu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-28T15:25:05Z","title":"AngelSpec: Towards Real-World High Performance Inference with Speculative Decoding"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.25852","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":"sha256:8f085f5ba2e600b6f50c5cde40f44a81c81cbe4534da02638f770425f3730c40","target":"record","created_at":"2026-07-29T01:26:06Z","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":"aef2a0a625ec79777aa5a04917b2ba03f969ed1e599bbb3732722c22c666bbd9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-28T15:25:05Z","title_canon_sha256":"3b2355ddcdc1ece1d22de78ea9bb1b1d766856c44ae4704e5ffa463bd093c6f6"},"schema_version":"1.0","source":{"id":"2607.25852","kind":"arxiv","version":1}},"canonical_sha256":"a8c585375d86108942618df73aa954cb42406a8cec42e0616dfb5da0836bf5a8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a8c585375d86108942618df73aa954cb42406a8cec42e0616dfb5da0836bf5a8","first_computed_at":"2026-07-29T01:26:06.712832Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-29T01:26:06.712832Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"LklIFNUkCppVEJ3SaTUyZyNr0TPtGlnI6FbE7bxpix6nfTfEGqlGzD/Oo9xu3/SZUWWPE7Y4QL5xQwRT/7DHDQ==","signature_status":"signed_v1","signed_at":"2026-07-29T01:26:06.713670Z","signed_message":"canonical_sha256_bytes"},"source_id":"2607.25852","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8f085f5ba2e600b6f50c5cde40f44a81c81cbe4534da02638f770425f3730c40","sha256:134f3a4ae2114b75e9ccda163e4d9d7dc304fc967b4894f11ed3afca7ec8a049"],"state_sha256":"cde9b02ebf2cc8a2263485e34436e6d8b63d2a254df9b035ced27f50537415e4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4e4fBPnABv2V6HQy/i8vnEAZWIMgh26qRfya1Fd7JM+BQvaWcRHjNPN7+lsm/tqV+eCTqo1reFO3DeRXTJtSDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-21T08:27:47.619498Z","bundle_sha256":"bad2c81e22b13ad62138254dc7e57d85b3641676c4caa33e4ae6abdc90fad414"}}