{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:24N2IJF6AHURMBX33MFFCBX5QJ","short_pith_number":"pith:24N2IJF6","canonical_record":{"source":{"id":"2603.22216","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-03-23T17:13:22Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"8a952f58ce043a36f795c3bc05fdd7dd2928978776fbb32a698669cf187a0cd7","abstract_canon_sha256":"276336514da028ff2805fc1605b0c2e908328d4f51269a6439b9ba777cca1997"},"schema_version":"1.0"},"canonical_sha256":"d71ba424be01e91606fbdb0a5106fd82704af6a739cd71a646f293a729911e7b","source":{"kind":"arxiv","id":"2603.22216","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2603.22216","created_at":"2026-07-24T01:24:09Z"},{"alias_kind":"arxiv_version","alias_value":"2603.22216v2","created_at":"2026-07-24T01:24:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.22216","created_at":"2026-07-24T01:24:09Z"},{"alias_kind":"pith_short_12","alias_value":"24N2IJF6AHUR","created_at":"2026-07-24T01:24:09Z"},{"alias_kind":"pith_short_16","alias_value":"24N2IJF6AHURMBX3","created_at":"2026-07-24T01:24:09Z"},{"alias_kind":"pith_short_8","alias_value":"24N2IJF6","created_at":"2026-07-24T01:24:09Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:24N2IJF6AHURMBX33MFFCBX5QJ","target":"record","payload":{"canonical_record":{"source":{"id":"2603.22216","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-03-23T17:13:22Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"8a952f58ce043a36f795c3bc05fdd7dd2928978776fbb32a698669cf187a0cd7","abstract_canon_sha256":"276336514da028ff2805fc1605b0c2e908328d4f51269a6439b9ba777cca1997"},"schema_version":"1.0"},"canonical_sha256":"d71ba424be01e91606fbdb0a5106fd82704af6a739cd71a646f293a729911e7b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-24T01:24:09.977143Z","signature_b64":"jmJSP6ghBUAokSHZqJN1FTj4aNr/d9jaZx+MNzV8SO6U+WcnsHJ2Up7RKxlFvm1JEA5SDQiEVUM6jcM3skVBCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d71ba424be01e91606fbdb0a5106fd82704af6a739cd71a646f293a729911e7b","last_reissued_at":"2026-07-24T01:24:09.975995Z","signature_status":"signed_v1","first_computed_at":"2026-07-24T01:24:09.975995Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2603.22216","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":"2026-07-24T01:24:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"g7SD+wV9mVt72PGPl2h0sGmMRqlIZqrgdhCT5mERcbuj+euQ1MS24ye/wD1Jj7EXgr0EXCwqinK66yVuazfHCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T03:28:07.631574Z"},"content_sha256":"acdf2d64012414f060adde44ea6f711f9566fef4d9c8feaae4aaa786575f0baf","schema_version":"1.0","event_id":"sha256:acdf2d64012414f060adde44ea6f711f9566fef4d9c8feaae4aaa786575f0baf"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:24N2IJF6AHURMBX33MFFCBX5QJ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Gumbel Distillation for Parallel Text Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Bo Liu, Chi Zhang, Qiang Liu, Xixi Hu","submitted_at":"2026-03-23T17:13:22Z","abstract_excerpt":"The slow, sequential nature of autoregressive (AR) language models has driven the adoption of parallel decoding methods. However, these non-AR models often sacrifice generation quality as they struggle to model the complex joint distribution of token sequences. To narrow this performance gap, we introduce Gumbel Distillation, a novel distillation technique that enables parallel decoders to learn this distribution effectively. Our method leverages the Gumbel-Max trick to create a deterministic mapping from a latent Gumbel noise space to the output tokens of a high-performing AR teacher. As a mo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.22216","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/2603.22216/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":"2026-07-24T01:24:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CLwJvZ+XQEd9vO7Al4XbIauXoqt2zHYxUMK2po0H6BfSSothSZ0WA1955OO0KH83br7AffsMDo9K4ASMQjjFBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T03:28:07.632181Z"},"content_sha256":"cbe824fc7a5fe5a50e9b8949dccb53faa2035ca2fb6aac639e30967932173e14","schema_version":"1.0","event_id":"sha256:cbe824fc7a5fe5a50e9b8949dccb53faa2035ca2fb6aac639e30967932173e14"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/24N2IJF6AHURMBX33MFFCBX5QJ/bundle.json","state_url":"https://pith.science/pith/24N2IJF6AHURMBX33MFFCBX5QJ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/24N2IJF6AHURMBX33MFFCBX5QJ/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":"2026-08-04T03:28:07Z","links":{"resolver":"https://pith.science/pith/24N2IJF6AHURMBX33MFFCBX5QJ","bundle":"https://pith.science/pith/24N2IJF6AHURMBX33MFFCBX5QJ/bundle.json","state":"https://pith.science/pith/24N2IJF6AHURMBX33MFFCBX5QJ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/24N2IJF6AHURMBX33MFFCBX5QJ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:24N2IJF6AHURMBX33MFFCBX5QJ","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":"276336514da028ff2805fc1605b0c2e908328d4f51269a6439b9ba777cca1997","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-03-23T17:13:22Z","title_canon_sha256":"8a952f58ce043a36f795c3bc05fdd7dd2928978776fbb32a698669cf187a0cd7"},"schema_version":"1.0","source":{"id":"2603.22216","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2603.22216","created_at":"2026-07-24T01:24:09Z"},{"alias_kind":"arxiv_version","alias_value":"2603.22216v2","created_at":"2026-07-24T01:24:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.22216","created_at":"2026-07-24T01:24:09Z"},{"alias_kind":"pith_short_12","alias_value":"24N2IJF6AHUR","created_at":"2026-07-24T01:24:09Z"},{"alias_kind":"pith_short_16","alias_value":"24N2IJF6AHURMBX3","created_at":"2026-07-24T01:24:09Z"},{"alias_kind":"pith_short_8","alias_value":"24N2IJF6","created_at":"2026-07-24T01:24:09Z"}],"graph_snapshots":[{"event_id":"sha256:cbe824fc7a5fe5a50e9b8949dccb53faa2035ca2fb6aac639e30967932173e14","target":"graph","created_at":"2026-07-24T01:24:09Z","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/2603.22216/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The slow, sequential nature of autoregressive (AR) language models has driven the adoption of parallel decoding methods. However, these non-AR models often sacrifice generation quality as they struggle to model the complex joint distribution of token sequences. To narrow this performance gap, we introduce Gumbel Distillation, a novel distillation technique that enables parallel decoders to learn this distribution effectively. Our method leverages the Gumbel-Max trick to create a deterministic mapping from a latent Gumbel noise space to the output tokens of a high-performing AR teacher. As a mo","authors_text":"Bo Liu, Chi Zhang, Qiang Liu, Xixi Hu","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-03-23T17:13:22Z","title":"Gumbel Distillation for Parallel Text Generation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.22216","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:acdf2d64012414f060adde44ea6f711f9566fef4d9c8feaae4aaa786575f0baf","target":"record","created_at":"2026-07-24T01:24:09Z","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":"276336514da028ff2805fc1605b0c2e908328d4f51269a6439b9ba777cca1997","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-03-23T17:13:22Z","title_canon_sha256":"8a952f58ce043a36f795c3bc05fdd7dd2928978776fbb32a698669cf187a0cd7"},"schema_version":"1.0","source":{"id":"2603.22216","kind":"arxiv","version":2}},"canonical_sha256":"d71ba424be01e91606fbdb0a5106fd82704af6a739cd71a646f293a729911e7b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d71ba424be01e91606fbdb0a5106fd82704af6a739cd71a646f293a729911e7b","first_computed_at":"2026-07-24T01:24:09.975995Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-24T01:24:09.975995Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"jmJSP6ghBUAokSHZqJN1FTj4aNr/d9jaZx+MNzV8SO6U+WcnsHJ2Up7RKxlFvm1JEA5SDQiEVUM6jcM3skVBCg==","signature_status":"signed_v1","signed_at":"2026-07-24T01:24:09.977143Z","signed_message":"canonical_sha256_bytes"},"source_id":"2603.22216","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:acdf2d64012414f060adde44ea6f711f9566fef4d9c8feaae4aaa786575f0baf","sha256:cbe824fc7a5fe5a50e9b8949dccb53faa2035ca2fb6aac639e30967932173e14"],"state_sha256":"04e7e6f2d8e77bac2cf2ab08a70a41b9ff534a22f6b9d1e3e185ff790d0201e7"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wOvpHeTcbiUKkacQE2Um6kQtnbbHQqH89c+htevGJ1ToogAUNBIT4zopBGvlVQszYSTQ7co9fZhNci5fjTPkCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T03:28:07.637047Z","bundle_sha256":"ad72da3986001291ff94c416ab69e4454787a04be9202011ab45b990431ec812"}}