{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:E2HVWNUZYUYA7J5MUDGLQ6VGYZ","short_pith_number":"pith:E2HVWNUZ","schema_version":"1.0","canonical_sha256":"268f5b3699c5300fa7aca0ccb87aa6c6555f9eb26cd0f82538370931922e21d0","source":{"kind":"arxiv","id":"2606.00091","version":1},"attestation_state":"computed","paper":{"title":"DLLM-JEPA: Joint Embedding Predictive Architectures for Masked Diffusion Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Sangdae Nam","submitted_at":"2026-05-24T16:14:54Z","abstract_excerpt":"Joint Embedding Predictive Architectures (JEPAs) have reshaped self-supervised representation learning in vision. The recent LLM-JEPA ported JEPA to autoregressive language models but inherited two steep costs from the causal-attention substrate: it demands explicit multi-view data (e.g., text-code pairs), and it requires two gradient-carrying forward passes per step. We introduce DLLM-JEPA, which pairs JEPA with masked-diffusion language models to eliminate both costs at once. The bidirectional attention of diffusion models yields two semantically distinct views of the same input via differen"},"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":"2606.00091","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-24T16:14:54Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"037d4b12eda4b9b01bca269038b7205e40b2b5a005985264c3344c95894f39e3","abstract_canon_sha256":"810556f45e0d7b2bff7315528f4e00a76c9382cf42496feee3f17f7ef13405b6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T00:03:14.744077Z","signature_b64":"1frGlqXVs7yQbY/HXwuF1PuAFggn/bMCLjgvx18t8E6xF5A5qricSYJpwPsKxcxw7Pk5zxir5cpyuoWuSgaCCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"268f5b3699c5300fa7aca0ccb87aa6c6555f9eb26cd0f82538370931922e21d0","last_reissued_at":"2026-06-02T00:03:14.743622Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T00:03:14.743622Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DLLM-JEPA: Joint Embedding Predictive Architectures for Masked Diffusion Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Sangdae Nam","submitted_at":"2026-05-24T16:14:54Z","abstract_excerpt":"Joint Embedding Predictive Architectures (JEPAs) have reshaped self-supervised representation learning in vision. The recent LLM-JEPA ported JEPA to autoregressive language models but inherited two steep costs from the causal-attention substrate: it demands explicit multi-view data (e.g., text-code pairs), and it requires two gradient-carrying forward passes per step. We introduce DLLM-JEPA, which pairs JEPA with masked-diffusion language models to eliminate both costs at once. The bidirectional attention of diffusion models yields two semantically distinct views of the same input via differen"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.00091","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/2606.00091/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":"2606.00091","created_at":"2026-06-02T00:03:14.743688+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.00091v1","created_at":"2026-06-02T00:03:14.743688+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.00091","created_at":"2026-06-02T00:03:14.743688+00:00"},{"alias_kind":"pith_short_12","alias_value":"E2HVWNUZYUYA","created_at":"2026-06-02T00:03:14.743688+00:00"},{"alias_kind":"pith_short_16","alias_value":"E2HVWNUZYUYA7J5M","created_at":"2026-06-02T00:03:14.743688+00:00"},{"alias_kind":"pith_short_8","alias_value":"E2HVWNUZ","created_at":"2026-06-02T00:03:14.743688+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/E2HVWNUZYUYA7J5MUDGLQ6VGYZ","json":"https://pith.science/pith/E2HVWNUZYUYA7J5MUDGLQ6VGYZ.json","graph_json":"https://pith.science/api/pith-number/E2HVWNUZYUYA7J5MUDGLQ6VGYZ/graph.json","events_json":"https://pith.science/api/pith-number/E2HVWNUZYUYA7J5MUDGLQ6VGYZ/events.json","paper":"https://pith.science/paper/E2HVWNUZ"},"agent_actions":{"view_html":"https://pith.science/pith/E2HVWNUZYUYA7J5MUDGLQ6VGYZ","download_json":"https://pith.science/pith/E2HVWNUZYUYA7J5MUDGLQ6VGYZ.json","view_paper":"https://pith.science/paper/E2HVWNUZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.00091&json=true","fetch_graph":"https://pith.science/api/pith-number/E2HVWNUZYUYA7J5MUDGLQ6VGYZ/graph.json","fetch_events":"https://pith.science/api/pith-number/E2HVWNUZYUYA7J5MUDGLQ6VGYZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/E2HVWNUZYUYA7J5MUDGLQ6VGYZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/E2HVWNUZYUYA7J5MUDGLQ6VGYZ/action/storage_attestation","attest_author":"https://pith.science/pith/E2HVWNUZYUYA7J5MUDGLQ6VGYZ/action/author_attestation","sign_citation":"https://pith.science/pith/E2HVWNUZYUYA7J5MUDGLQ6VGYZ/action/citation_signature","submit_replication":"https://pith.science/pith/E2HVWNUZYUYA7J5MUDGLQ6VGYZ/action/replication_record"}},"created_at":"2026-06-02T00:03:14.743688+00:00","updated_at":"2026-06-02T00:03:14.743688+00:00"}