{"paper":{"title":"Prefilling-dLLM: Predictive Prefilling for Long-Context Inference in Diffusion Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chaofan Tao, Chengyue Wu, Chenyang Zhao, Jing Xiong, Ngai Wong, Qi Han, Shansan Gong, Yunta Hsieh","submitted_at":"2026-06-09T08:06:22Z","abstract_excerpt":"Diffusion large language models (dLLMs) re-encode the entire prefix at every denoising step, causing recomputation that scales\n  quadratically with context length and becomes prohibitive for long-context scenarios. We propose Prefilling-dLLM, a training-free\n  prefill-decode disaggregation framework for dLLMs that partitions the prefix into N chunks, caches their KV representations once,\n  and selects the top-K most relevant chunks with intra-chunk token sparsity for decoding, showing that sparse prefilling can\n  outperform dense attention while reducing per-step complexity from quadratic in t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.10537","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/2606.10537/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"}