{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:747C7OLEYMLBG7XMQ72ID2KVHN","short_pith_number":"pith:747C7OLE","schema_version":"1.0","canonical_sha256":"ff3e2fb964c316137eec87f481e9553b6ae926add2effa23d8640bf6c001e9a2","source":{"kind":"arxiv","id":"2606.01024","version":1},"attestation_state":"computed","paper":{"title":"DSL-LLaDA: Scaling Continuous Denoising to 8B Masked Diffusion LMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Greg Ver Steeg, Hui Liu, Longxuan Yu, Rob Brekelmans, Siheng Xiong, Yue Dong, Yu Fu, Yunshu Wu","submitted_at":"2026-05-31T05:27:01Z","abstract_excerpt":"Discrete Masked diffusion language models generate text by iterative parallel decoding, but few-step decoding suffers from a tradeoff between length and quality: with a fixed step budget, standard methods can generate a short, high-quality output, or they can produce long but repetitive text. Continuous denoising can sidestep this tradeoff by evolving all positions jointly in embedding space, but building such a model from scratch at scale remains an open problem. We show that a pretrained masked DLM can instead be lightly adapted to support continuous embedding-space denoising. Starting from "},"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":"2606.01024","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-31T05:27:01Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e286c27decfcfd0d54db7e0b89c64c7b732b5492c3ef029dc684da119a10f1da","abstract_canon_sha256":"1a6f1d82aedb75ff83470cda2faf4fe0688a0196fdb8799ea0727df03254f75f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:04:18.750819Z","signature_b64":"orOGH45AV+IrJM+n8T+R3/etQ1m1ax5g+uFvLdo0fZRzIgLfibQIfDLcPHiyNkJP+r1VIXzBJ12Ro7vvvTmQAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ff3e2fb964c316137eec87f481e9553b6ae926add2effa23d8640bf6c001e9a2","last_reissued_at":"2026-06-02T01:04:18.750364Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:04:18.750364Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DSL-LLaDA: Scaling Continuous Denoising to 8B Masked Diffusion LMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Greg Ver Steeg, Hui Liu, Longxuan Yu, Rob Brekelmans, Siheng Xiong, Yue Dong, Yu Fu, Yunshu Wu","submitted_at":"2026-05-31T05:27:01Z","abstract_excerpt":"Discrete Masked diffusion language models generate text by iterative parallel decoding, but few-step decoding suffers from a tradeoff between length and quality: with a fixed step budget, standard methods can generate a short, high-quality output, or they can produce long but repetitive text. Continuous denoising can sidestep this tradeoff by evolving all positions jointly in embedding space, but building such a model from scratch at scale remains an open problem. We show that a pretrained masked DLM can instead be lightly adapted to support continuous embedding-space denoising. Starting from "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.01024","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/2606.01024/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":"2606.01024","created_at":"2026-06-02T01:04:18.750432+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.01024v1","created_at":"2026-06-02T01:04:18.750432+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.01024","created_at":"2026-06-02T01:04:18.750432+00:00"},{"alias_kind":"pith_short_12","alias_value":"747C7OLEYMLB","created_at":"2026-06-02T01:04:18.750432+00:00"},{"alias_kind":"pith_short_16","alias_value":"747C7OLEYMLBG7XM","created_at":"2026-06-02T01:04:18.750432+00:00"},{"alias_kind":"pith_short_8","alias_value":"747C7OLE","created_at":"2026-06-02T01:04:18.750432+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/747C7OLEYMLBG7XMQ72ID2KVHN","json":"https://pith.science/pith/747C7OLEYMLBG7XMQ72ID2KVHN.json","graph_json":"https://pith.science/api/pith-number/747C7OLEYMLBG7XMQ72ID2KVHN/graph.json","events_json":"https://pith.science/api/pith-number/747C7OLEYMLBG7XMQ72ID2KVHN/events.json","paper":"https://pith.science/paper/747C7OLE"},"agent_actions":{"view_html":"https://pith.science/pith/747C7OLEYMLBG7XMQ72ID2KVHN","download_json":"https://pith.science/pith/747C7OLEYMLBG7XMQ72ID2KVHN.json","view_paper":"https://pith.science/paper/747C7OLE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.01024&json=true","fetch_graph":"https://pith.science/api/pith-number/747C7OLEYMLBG7XMQ72ID2KVHN/graph.json","fetch_events":"https://pith.science/api/pith-number/747C7OLEYMLBG7XMQ72ID2KVHN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/747C7OLEYMLBG7XMQ72ID2KVHN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/747C7OLEYMLBG7XMQ72ID2KVHN/action/storage_attestation","attest_author":"https://pith.science/pith/747C7OLEYMLBG7XMQ72ID2KVHN/action/author_attestation","sign_citation":"https://pith.science/pith/747C7OLEYMLBG7XMQ72ID2KVHN/action/citation_signature","submit_replication":"https://pith.science/pith/747C7OLEYMLBG7XMQ72ID2KVHN/action/replication_record"}},"created_at":"2026-06-02T01:04:18.750432+00:00","updated_at":"2026-06-02T01:04:18.750432+00:00"}