{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:N72JVDRC53JLY7BLYXXIUQQS72","short_pith_number":"pith:N72JVDRC","schema_version":"1.0","canonical_sha256":"6ff49a8e22eed2bc7c2bc5ee8a4212febbd7d8a8e5bdae9eb7b0b82abaa9381e","source":{"kind":"arxiv","id":"2605.27387","version":1},"attestation_state":"computed","paper":{"title":"From AR to Diffusion: Efficiently Adapting Large Language Models with Strictly Causal and Elastic Horizons","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Lefei Zhang, Teng Xiao, Xiangyu Ma, Zuchao Li","submitted_at":"2026-04-11T13:18:27Z","abstract_excerpt":"Diffusion models promise efficient parallel text generation but rely on bidirectional attention, creating a structural mismatch with pre-trained Autoregressive (AR) models. This incompatibility precludes reusing robust AR priors, necessitating prohibitive pre-training from scratch. To bridge this gap, we propose FLUID, a framework that efficiently adapts AR backbones to the diffusion paradigm. By enforcing Strictly Causal Alignment, FLUID enables seamless initialization from standard GPT-style checkpoints, circumventing the need for massive pre-training. Furthermore, we introduce Elastic Horiz"},"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":"2605.27387","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-04-11T13:18:27Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"d71030c7a0c6af6cec3960a78ffec924b25fef40cac971cdff15519ddd11caf9","abstract_canon_sha256":"77aacf7ed23e819f8918492241fbfa447f376bf8f5e18c5a5c5adddfe36128d1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T00:05:16.311807Z","signature_b64":"BRD0Fx9pOOPFSVX6Loq9qh8NWTFiKdw1ee1jEm+UteIiyaMaZEl9puAfrO7Hg/K1n1Qq/rktnvMxpt1l/ggXBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6ff49a8e22eed2bc7c2bc5ee8a4212febbd7d8a8e5bdae9eb7b0b82abaa9381e","last_reissued_at":"2026-05-28T00:05:16.311160Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T00:05:16.311160Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From AR to Diffusion: Efficiently Adapting Large Language Models with Strictly Causal and Elastic Horizons","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Lefei Zhang, Teng Xiao, Xiangyu Ma, Zuchao Li","submitted_at":"2026-04-11T13:18:27Z","abstract_excerpt":"Diffusion models promise efficient parallel text generation but rely on bidirectional attention, creating a structural mismatch with pre-trained Autoregressive (AR) models. This incompatibility precludes reusing robust AR priors, necessitating prohibitive pre-training from scratch. To bridge this gap, we propose FLUID, a framework that efficiently adapts AR backbones to the diffusion paradigm. By enforcing Strictly Causal Alignment, FLUID enables seamless initialization from standard GPT-style checkpoints, circumventing the need for massive pre-training. Furthermore, we introduce Elastic Horiz"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.27387","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/2605.27387/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":"2605.27387","created_at":"2026-05-28T00:05:16.311231+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.27387v1","created_at":"2026-05-28T00:05:16.311231+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.27387","created_at":"2026-05-28T00:05:16.311231+00:00"},{"alias_kind":"pith_short_12","alias_value":"N72JVDRC53JL","created_at":"2026-05-28T00:05:16.311231+00:00"},{"alias_kind":"pith_short_16","alias_value":"N72JVDRC53JLY7BL","created_at":"2026-05-28T00:05:16.311231+00:00"},{"alias_kind":"pith_short_8","alias_value":"N72JVDRC","created_at":"2026-05-28T00:05:16.311231+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/N72JVDRC53JLY7BLYXXIUQQS72","json":"https://pith.science/pith/N72JVDRC53JLY7BLYXXIUQQS72.json","graph_json":"https://pith.science/api/pith-number/N72JVDRC53JLY7BLYXXIUQQS72/graph.json","events_json":"https://pith.science/api/pith-number/N72JVDRC53JLY7BLYXXIUQQS72/events.json","paper":"https://pith.science/paper/N72JVDRC"},"agent_actions":{"view_html":"https://pith.science/pith/N72JVDRC53JLY7BLYXXIUQQS72","download_json":"https://pith.science/pith/N72JVDRC53JLY7BLYXXIUQQS72.json","view_paper":"https://pith.science/paper/N72JVDRC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.27387&json=true","fetch_graph":"https://pith.science/api/pith-number/N72JVDRC53JLY7BLYXXIUQQS72/graph.json","fetch_events":"https://pith.science/api/pith-number/N72JVDRC53JLY7BLYXXIUQQS72/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/N72JVDRC53JLY7BLYXXIUQQS72/action/timestamp_anchor","attest_storage":"https://pith.science/pith/N72JVDRC53JLY7BLYXXIUQQS72/action/storage_attestation","attest_author":"https://pith.science/pith/N72JVDRC53JLY7BLYXXIUQQS72/action/author_attestation","sign_citation":"https://pith.science/pith/N72JVDRC53JLY7BLYXXIUQQS72/action/citation_signature","submit_replication":"https://pith.science/pith/N72JVDRC53JLY7BLYXXIUQQS72/action/replication_record"}},"created_at":"2026-05-28T00:05:16.311231+00:00","updated_at":"2026-05-28T00:05:16.311231+00:00"}