{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:46GP7IXFFJGQUCVH7TCI7YW5DT","short_pith_number":"pith:46GP7IXF","schema_version":"1.0","canonical_sha256":"e78cffa2e52a4d0a0aa7fcc48fe2dd1cf00ea716fc4ca5049cd9cb97fecbff05","source":{"kind":"arxiv","id":"2508.14090","version":2},"attestation_state":"computed","paper":{"title":"DLLMQuant: Quantizing Diffusion-based Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chen Xu, Dawei Yang","submitted_at":"2025-08-14T09:30:17Z","abstract_excerpt":"Diffusion-based large language models (DLLMs) have shown promise for non-autoregressive text generation, but their deployment is constrained by large model sizes and heavy computational costs. Post-training quantization (PTQ), a widely used method for compressing and accelerating Large Language Models (LLMs), suffers from severe accuracy degradation and reduced generalization performance when directly applied to DLLMs (e.g., AWQ suffers a 16% accuracy drop on LLADA under W4A4). This paper explores how DLLMs' key mechanisms - dynamic masking, iterative generation, bidirectional attention - clas"},"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":"2508.14090","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-08-14T09:30:17Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"dfcaad016fef7a624231e5c19c77969df6447e93340a2d4909db8dc9faca901a","abstract_canon_sha256":"a43f0566042ef7b98d8ca2940a6ee431815ef76f211e97905fb76ce7c626f3cc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:59:20.157650Z","signature_b64":"UIWofllV4bz6rOHktU1RGM4klMIePz49gK3N076DiF6VLVn7kvOYYz8LujKgfdtNtGTTw/ZlXs2U7utJN1P3Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e78cffa2e52a4d0a0aa7fcc48fe2dd1cf00ea716fc4ca5049cd9cb97fecbff05","last_reissued_at":"2026-07-05T11:59:20.157185Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:59:20.157185Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DLLMQuant: Quantizing Diffusion-based Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chen Xu, Dawei Yang","submitted_at":"2025-08-14T09:30:17Z","abstract_excerpt":"Diffusion-based large language models (DLLMs) have shown promise for non-autoregressive text generation, but their deployment is constrained by large model sizes and heavy computational costs. Post-training quantization (PTQ), a widely used method for compressing and accelerating Large Language Models (LLMs), suffers from severe accuracy degradation and reduced generalization performance when directly applied to DLLMs (e.g., AWQ suffers a 16% accuracy drop on LLADA under W4A4). This paper explores how DLLMs' key mechanisms - dynamic masking, iterative generation, bidirectional attention - clas"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.14090","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/2508.14090/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":"2508.14090","created_at":"2026-07-05T11:59:20.157246+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.14090v2","created_at":"2026-07-05T11:59:20.157246+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.14090","created_at":"2026-07-05T11:59:20.157246+00:00"},{"alias_kind":"pith_short_12","alias_value":"46GP7IXFFJGQ","created_at":"2026-07-05T11:59:20.157246+00:00"},{"alias_kind":"pith_short_16","alias_value":"46GP7IXFFJGQUCVH","created_at":"2026-07-05T11:59:20.157246+00:00"},{"alias_kind":"pith_short_8","alias_value":"46GP7IXF","created_at":"2026-07-05T11:59:20.157246+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.08302","citing_title":"DMax: Aggressive Parallel Decoding for dLLMs","ref_index":88,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18995","citing_title":"$R^2$-dLLM: Accelerating Diffusion Large Language Models via Spatio-Temporal Redundancy Reduction","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08302","citing_title":"DMax: Aggressive Parallel Decoding for dLLMs","ref_index":88,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/46GP7IXFFJGQUCVH7TCI7YW5DT","json":"https://pith.science/pith/46GP7IXFFJGQUCVH7TCI7YW5DT.json","graph_json":"https://pith.science/api/pith-number/46GP7IXFFJGQUCVH7TCI7YW5DT/graph.json","events_json":"https://pith.science/api/pith-number/46GP7IXFFJGQUCVH7TCI7YW5DT/events.json","paper":"https://pith.science/paper/46GP7IXF"},"agent_actions":{"view_html":"https://pith.science/pith/46GP7IXFFJGQUCVH7TCI7YW5DT","download_json":"https://pith.science/pith/46GP7IXFFJGQUCVH7TCI7YW5DT.json","view_paper":"https://pith.science/paper/46GP7IXF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.14090&json=true","fetch_graph":"https://pith.science/api/pith-number/46GP7IXFFJGQUCVH7TCI7YW5DT/graph.json","fetch_events":"https://pith.science/api/pith-number/46GP7IXFFJGQUCVH7TCI7YW5DT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/46GP7IXFFJGQUCVH7TCI7YW5DT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/46GP7IXFFJGQUCVH7TCI7YW5DT/action/storage_attestation","attest_author":"https://pith.science/pith/46GP7IXFFJGQUCVH7TCI7YW5DT/action/author_attestation","sign_citation":"https://pith.science/pith/46GP7IXFFJGQUCVH7TCI7YW5DT/action/citation_signature","submit_replication":"https://pith.science/pith/46GP7IXFFJGQUCVH7TCI7YW5DT/action/replication_record"}},"created_at":"2026-07-05T11:59:20.157246+00:00","updated_at":"2026-07-05T11:59:20.157246+00:00"}