{"_canonicalization":{"envelope_id":"axm_ + sha256(envelope minus {signature, axiom_id, anchors})","envelope_signature":"ed25519(envelope minus {signature, axiom_id})","json":"sort_keys=True, separators=(',',':'), ensure_ascii=False, allow_nan=False, utf-8","leaf_hash":"sha256(0x00 || canonical_json(envelope_full))","seal_signature":"ed25519(seal minus {signature, sig_algorithm})"},"axiom_id":"axm_879c984a218d449992789c81405b9a85b416e6f330dcb9617bdfdcee19514644","bitcoin_anchor":{"bitcoin_attestations":[],"calendar_attestations":[],"ots_url":"","stamped_at":"","status":"pending_next_stamp"},"envelope":{"anchors":[{"chain":"crovia.axiom_graph","height":0,"merkle_proof":"spider_vendor_press_v1","root_at_anchor":"spider_vendor_press_v1"}],"axiom_id":"axm_879c984a218d449992789c81405b9a85b416e6f330dcb9617bdfdcee19514644","axiom_type":"AX.OBS","body":{"axiom_subtype":"news.vendor_press.v1","category":"news","fingerprint":"fc9a966fc64f8f8920fffbe3c116f0eb1e02a491404488b25b39e701f126898c","published":"Tue, 21 Jul 2026 00:00:00 -0400","receipt_hash":"fc9a966fc64f8f8920fffbe3c116f0eb1e02a491404488b25b39e701f126898c","schema":"spider.news.vendor_press.v1","spider":"vendor_press","spider_record":{"axiom_subtype":"news.vendor_press.v1","category":"news","decision_hint":"POSITIVE","envelope_target":"AX.OBS","fingerprint":"fc9a966fc64f8f8920fffbe3c116f0eb1e02a491404488b25b39e701f126898c","observed_at":"2026-07-21T04:43:35.036805Z","parent_run_hash":"03e944014de2697434479833d15ea9303e014945afc230ecc7f207824493b589","published":"Tue, 21 Jul 2026 00:00:00 -0400","runtime_version":"0.1.0","schema":"spider.news.vendor_press.v1","source_status":200,"source_url":"https://export.arxiv.org/rss/cs.AI","spider":"vendor_press","summary_excerpt":"arXiv:2607.16872v1 Announce Type: cross \nAbstract: Diffusion large language models (dLLMs) are a promising alternative to autoregressive generation. However, reasoning-oriented post-training for dLLMs remains challenging. Supervised fine-tuning (SFT) for dLLMs requires dense but often off-policy masked states, while reinforcement learning (RL) relies on sparse rewards or value modeling. This paper proposes \\textbf{trace-based on-policy distillation (TOPD)}, a teacher-supervised framework that transfers reasoning ability to a target dLLM without reward estimation. The key idea is to supervise a dLLM on its own denoising trajectory, focusing on the trace-aligned token decisions that form the final response. Specifically, TOPD samples on-policy diffusion trajectories from the target dLLM, obtains teacher token distributions from a teacher model on the corresponding partially denoised states, and updates the target dLLM with a token-level Reverse Kullback-Leibler (Reverse-KL) objective. Th","title":"Trace-Based On-Policy Distillation for Masked Diffusion Language Models","url":"https://arxiv.org/abs/2607.16872","vendor":"arxiv_cs_ai"},"summary":"arXiv:2607.16872v1 Announce Type: cross \nAbstract: Diffusion large language models (dLLMs) are a promising alternative to autoregressive generation. However, reasoning-oriented post-training for dLLMs remains challenging. Supervised fine-tuning (SFT) for dLLMs requires dense but often off-policy masked states, while reinforcement learning (RL) relies on sparse rewards or value modeling. This paper proposes \\textbf{trace-based on-policy distillation (TOPD)}, a teacher-supervised framework that transfers reasoning ability to a target dLLM without reward estimation. The key idea is to supervise a dLLM on its own denoising trajectory, focusing on the trace-aligned token decisions that form the final response. Specifically, TOPD samples on-policy diffusion trajectories from the target dLLM, obtains teacher token distributions from a teacher model on the corresponding partially denoised states, and updates the target dLLM with a token-level Reverse Kullback-Leibler (Reverse-KL) objective. Th","title":"Trace-Based On-Policy Distillation for Masked Diffusion Language Models","vendor":"arxiv_cs_ai"},"confidence":{"method":"deterministic"},"decision":"POSITIVE","issued_at":"2026-07-21T04:43:35Z","notes":"Spider vendor_press (news) news.vendor_press.v1","object":{"captured_by":"crovia.spider.vendor_press","primary_source_url":"https://arxiv.org/abs/2607.16872"},"predecessors":[],"schema":"crovia.axiom.v1","signature":"ed25519:252da153c68b2f25384c1b49b0f396bf37c1e41389165810c6765ecd15c4bfceb98f7f6a97eb66432f11fc0077897a8ca2f3480e2656989a90812af198725805","signer":"crovia.substrate","subject":{"observed_at":"2026-07-21T04:43:35Z","source_collector":"spider:vendor_press","target_id":"https://arxiv.org/abs/2607.16872"},"tsa":{"authority":"crovia.substrate.bootstrap","rfc3161_token":"{\"kind\":\"crovia.bootstrap.tsa\",\"source_jsonl\":\"/opt/crovia/spider/data/news/vendor_press_v1.jsonl\",\"source_seal_merkle_root\":\"spider_vendor_press_v1\",\"upgrade_path\":\"Sessione H \\u2014 OpenTimestamps weekly anchor\"}"},"zk_mode":"clear","zk_proof":null},"ledger":{"leaf_hash":"47b5b22faae95fab04d956c40a9ed36c494ca97957246615d00af9c8b80517d0","leaf_index":336722,"ledger_path":"/opt/crovia/substrate/axiom_ledger.jsonl"},"merkle_proof":{"hash_alg":"sha256","leaf_prefix":"0x00","node_prefix":"0x01","odd_leaf_rule":"duplicate_last","path":[{"sibling":"69cae5cdad033b5fb0852963f0bdaf27b6af04aa7b21e3387bd3276010808aed","side":"right"},{"sibling":"8c944a4a1c6ea2780e316c964b9511af57c1926aedc58f4be5e627fe7d838869","side":"left"},{"sibling":"720a0f6801615269650aa96c357ce018261a961e10add94db1d9d922dac08efc","side":"right"},{"sibling":"32384d9a08bad59acbdfc76dcd0beff63b1b105daf7fce7a81c10f98b60e33f3","side":"right"},{"sibling":"f0061fcafbf8e6a8152bfbef71e4b13ad9956e1985fbb8157db2c1f947936e44","side":"left"},{"sibling":"a67d55fd22cd262abeda1495d0d2509734cbc60bb2980c0014022699d323696d","side":"right"},{"sibling":"c3453c4355287f8f6ea5b33a2f2d95f482fa4ede383b4407cc8d044e332ee8e5","side":"left"},{"sibling":"a4c41a614872a391d764732d35b746845ccfe3f2a51fe13e1f0e934d271e94f4","side":"right"},{"sibling":"bd233cee8c0876447824d86de33608c36e3a8e17dca0737155291ed768cf56d1","side":"left"},{"sibling":"7b927551b5db06b6571913b4e6792ffcce5291a3eca5a0df4a3b6e296271105f","side":"left"},{"sibling":"b77a0b5ae4607c8fe6ba73449d46b35076e3dedc0c82a2c65a05780d42a7bc2e","side":"right"},{"sibling":"9eb5077edfb3dc553857d4794b925bfce117e0f8a1d049af5d0dd9026b470eef","side":"right"},{"sibling":"414b1a70fd1dcb25489a194714b97492b066684b15d0b7a48a176c4b9b5bc713","side":"right"},{"sibling":"21d66dd41003813f710b7617944f1bfba3258658a5d3370c21cad8f9e945bc99","side":"left"},{"sibling":"613f015699131eb89bd755dee67133be95af25cf5f16c1c8ce4b99d963b8dd86","side":"right"},{"sibling":"a729b574b1135956436ded5eef1fe8f08014ff6a0729749d307ab1bca93fcdc9","side":"right"},{"sibling":"77025bcb374a7ad74f520643e20a8ae1205a7ee78507b0beb93117760f1c29d3","side":"left"},{"sibling":"4bf21052e085e8ac81f1dec1d2b310bd12bf948992de6177d12e9d2fda8d39f0","side":"right"},{"sibling":"1cecb7f447febd025aac272837c80de218aecc6485d2395a509b2a1f1b9c746e","side":"left"}]},"schema":"crovia.axiom_proof.v1","seal":{"first_collector_run_id":"","first_receipt_hash":"","jsonl_path":"/opt/crovia/substrate/axiom_ledger.jsonl","key_id":"430895f101d38164","last_collector_run_id":"","last_receipt_hash":"","leaf_count":337144,"merkle_root":"5e969cc01afa67e4dbe5d37b712cdb10f4aa1fd74404e02eab724cf487c8d6d9","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","run_id":"hourly_json_retrofit_20260721T053701Z","schema":"crovia.seal.v1","seal_family_version":"crovia-seal-family/1","seal_kind":"substrate_batch","sealed_at":"2026-07-21T05:38:38Z","sig_algorithm":"ed25519","signature":"5c0c1a8dd2793d787ccd5e49e8b4d70eed136352555518589f05c74be357fc171702e42a76c3a556d90e3d51ff36cb3d292aaac83c66566de7b943f318bda50c","signer_version":"1.1.0"},"trust_root":{"key_id":"430895f101d38164","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","signature_algorithm":"ed25519","url":"/registry/canon/TRUST_ROOT.md"},"verifier":{"spec":"/registry/canon/AXIOM_RECEIPT_v1.md","url":"/v/axm_879c984a218d449992789c81405b9a85b416e6f330dcb9617bdfdcee19514644"}}