{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:34TMSD6DHFBBWLM5NTF2ISU4I2","short_pith_number":"pith:34TMSD6D","schema_version":"1.0","canonical_sha256":"df26c90fc339421b2d9d6ccba44a9c46868c29cd7e8b6dca784b7f9a3a51d6ad","source":{"kind":"arxiv","id":"2605.26106","version":1},"attestation_state":"computed","paper":{"title":"Looped Diffusion Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chunsan Hong, Dongmin Park, Jongho Park, Jonghyun Lee, Sanghyun Lee, Seungryong Kim","submitted_at":"2026-05-25T17:58:24Z","abstract_excerpt":"Masked diffusion models (MDMs) have emerged as a promising alternative to autoregressive models for language modeling, yet the effective design of transformer architectures for MDMs remains underexplored. In this paper, we show that selectively looping the early-middle transformer layers significantly improves both training efficiency and model performance in MDMs. We call this approach LoopMDM(Looped Masked Diffusion Model), which brings two key benefits: looping layers at training-time yields a depth-scaling effect without adding parameters, while varying the number of loops at inference-tim"},"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":"2605.26106","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T17:58:24Z","cross_cats_sorted":[],"title_canon_sha256":"cb661905690d3a1a233ec9503b611d3246b9eaf883e9d868ca100f78936b04e3","abstract_canon_sha256":"0470a74322e614435eec77aeb68a7cc1012b523175286a0667e23c5a9b8d1ef1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T03:05:24.378642Z","signature_b64":"cAMPA7IDDzOWB9vq4N8vskGRYV/IbusKCwy00Vv/hXhY3114bqJxGdry3uVShxPrhYM9jjl1w5fatEh0BpGFAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"df26c90fc339421b2d9d6ccba44a9c46868c29cd7e8b6dca784b7f9a3a51d6ad","last_reissued_at":"2026-05-26T03:05:24.377586Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T03:05:24.377586Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Looped Diffusion Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chunsan Hong, Dongmin Park, Jongho Park, Jonghyun Lee, Sanghyun Lee, Seungryong Kim","submitted_at":"2026-05-25T17:58:24Z","abstract_excerpt":"Masked diffusion models (MDMs) have emerged as a promising alternative to autoregressive models for language modeling, yet the effective design of transformer architectures for MDMs remains underexplored. In this paper, we show that selectively looping the early-middle transformer layers significantly improves both training efficiency and model performance in MDMs. We call this approach LoopMDM(Looped Masked Diffusion Model), which brings two key benefits: looping layers at training-time yields a depth-scaling effect without adding parameters, while varying the number of loops at inference-tim"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.26106","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/2605.26106/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":"2605.26106","created_at":"2026-05-26T03:05:24.377700+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.26106v1","created_at":"2026-05-26T03:05:24.377700+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.26106","created_at":"2026-05-26T03:05:24.377700+00:00"},{"alias_kind":"pith_short_12","alias_value":"34TMSD6DHFBB","created_at":"2026-05-26T03:05:24.377700+00:00"},{"alias_kind":"pith_short_16","alias_value":"34TMSD6DHFBBWLM5","created_at":"2026-05-26T03:05:24.377700+00:00"},{"alias_kind":"pith_short_8","alias_value":"34TMSD6D","created_at":"2026-05-26T03:05:24.377700+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/34TMSD6DHFBBWLM5NTF2ISU4I2","json":"https://pith.science/pith/34TMSD6DHFBBWLM5NTF2ISU4I2.json","graph_json":"https://pith.science/api/pith-number/34TMSD6DHFBBWLM5NTF2ISU4I2/graph.json","events_json":"https://pith.science/api/pith-number/34TMSD6DHFBBWLM5NTF2ISU4I2/events.json","paper":"https://pith.science/paper/34TMSD6D"},"agent_actions":{"view_html":"https://pith.science/pith/34TMSD6DHFBBWLM5NTF2ISU4I2","download_json":"https://pith.science/pith/34TMSD6DHFBBWLM5NTF2ISU4I2.json","view_paper":"https://pith.science/paper/34TMSD6D","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.26106&json=true","fetch_graph":"https://pith.science/api/pith-number/34TMSD6DHFBBWLM5NTF2ISU4I2/graph.json","fetch_events":"https://pith.science/api/pith-number/34TMSD6DHFBBWLM5NTF2ISU4I2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/34TMSD6DHFBBWLM5NTF2ISU4I2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/34TMSD6DHFBBWLM5NTF2ISU4I2/action/storage_attestation","attest_author":"https://pith.science/pith/34TMSD6DHFBBWLM5NTF2ISU4I2/action/author_attestation","sign_citation":"https://pith.science/pith/34TMSD6DHFBBWLM5NTF2ISU4I2/action/citation_signature","submit_replication":"https://pith.science/pith/34TMSD6DHFBBWLM5NTF2ISU4I2/action/replication_record"}},"created_at":"2026-05-26T03:05:24.377700+00:00","updated_at":"2026-05-26T03:05:24.377700+00:00"}