mlsanityData to 14 Oct 2026

Fast-dLLM++: Fr'echet Profile Decoding for Faster Diffusion LLM Inference

arXiv 2606.02955 ↗First posted 1 Jun 2026cs.CL cs.AI cs.LGcited by 2

Not on the map. Why?

Abstract (arXiv)

Diffusion large language models promise parallel token generation, yet inference remains bottlenecked by deciding which masked tokens can be safely committed together. Fast-dLLM addressed this with KV caching and confidence-guided parallel decoding, but its decoding theory uses a homogeneous high-confidence assumption that effectively reduces each candidate set to its weakest selected token. We argue that this leaves speed on the table because real decoding steps exhibit heterogeneous confidence profiles. We propose Fast-dLLM++, a training-free extension that introduces emphFr'echet profile decoding: selecting parallel commit sets from the full sorted confidence profile rather than a single worst-case confidence. The resulting rule is a heterogeneous-confidence generalization of Fast-dLLM's factor selector and it recovers the previous rule exactly in the equal-confidence case and adds a provable heterogeneity bonus when the selected tokens have uneven confidences. Fast-dLLM++ leaves the model, diffusion process, and cache implementation entirely unchanged, making it a drop-in replacement for existing Fast-dLLM decoding. Experiments on GSM8K, MATH, HumanEval, and MBPP with the LLaDA-8B model show that the theoretical improvement translates directly into empirical gains: profile-aware selection improves the accuracy--throughput frontier by exploiting safe parallelism that weakest-token rules miss, achieving up to 37% higher throughput at comparable accuracy. Our code release is at https://github.com/Ringo-Star/FastdLLMplusplus.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 44 of its 51 references that are in the map. First posted on arXiv 1 Jun 2026.

Builds on

44 of 51 references.

References · 44

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  3. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
  4. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  5. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
  6. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  7. 2017On Calibration of Modern Neural Networkscited by 4,960
  8. 2021Program Synthesis with Large Language Modelscited by 3,637
  9. 2022Language Models (Mostly) Know What They Knowcited by 1,856
  10. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,719
  11. 2022Fast Inference from Transformers via Speculative Decodingcited by 1,653
  12. 2023MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contextscited by 1,614
32 more
  1. 2022Diffusion-LM Improves Controllable Text Generationcited by 1,027
  2. 2023Accelerating Large Language Model Decoding with Speculative Samplingcited by 1,006
  3. 2025Large Language Diffusion Modelscited by 912
  4. 2024Simple and Effective Masked Diffusion Language Modelscited by 848
  5. 2023Discrete Diffusion Modeling by Estimating the Ratios of the Data Distributioncited by 745
  6. 2024Simplified and Generalized Masked Diffusion for Discrete Datacited by 626
  7. 2024MathVerse: Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems?cited by 586
  8. 2021Argmax Flows and Multinomial Diffusion: Learning Categorical Distributionscited by 574
  9. 2025Dream 7B: Diffusion Large Language Modelscited by 544
  10. 2022A Continuous Time Framework for Discrete Denoising Modelscited by 451
  11. 2022DiffuSeq: Sequence to Sequence Text Generation with Diffusion Modelscited by 450
  12. 2020How Can We Know When Language Models Know? On the Calibration of Language Models for Question Answeringcited by 445
  13. 2025Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decodingcited by 420
  14. 2025Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Modelscited by 409
  15. 2018Blockwise Parallel Decoding for Deep Autoregressive Modelscited by 323
  16. 2020Calibration of Pre-trained Transformerscited by 315
  17. 2022DiffusionBERT: Improving Generative Masked Language Models with Diffusion Modelscited by 214
  18. 2022SSD-LM: Semi-autoregressive Simplex-based Diffusion Language Model for Text Generation and Modular Controlcited by 194
  19. 2022Score-based Continuous-time Discrete Diffusion Modelscited by 169
  20. 2025LLaDA-V: Large Language Diffusion Models with Visual Instruction Tuningcited by 148
  21. 2025Accelerated Sampling from Masked Diffusion Models via Entropy Bounded Unmaskingcited by 131
  22. 2025Accelerating Diffusion LLMs via Adaptive Parallel Decodingcited by 112
  23. 2025Dimple: Discrete Diffusion Multimodal Large Language Model with Parallel Decodingcited by 101
  24. 2024Discrete Copula Diffusioncited by 54
  25. 2025ParallelBench: Understanding the Trade-offs of Parallel Decoding in Diffusion LLMscited by 43
  26. 2025Parallel Sampling from Masked Diffusion Models via Conditional Independence Testingcited by 13
  27. 2026Parallelism and Generation Order in Masked Diffusion Language Models: Limits Today, Potential Tomorrowcited by 11
  28. 2026Why Diffusion Language Models Struggle with Truly Parallel (Non-Autoregressive) Decoding?cited by 11
  29. 2025Enabling Approximate Joint Sampling in Diffusion LMscited by 10
  30. 2025Generative or Discriminative? Revisiting Text Classification in the Era of Transformerscited by 9
  31. 2020Improved Inference of Gaussian Mixture Copula Model for Clustering and Reproducibility Analysis using Automatic Differentiationcited by 4
  32. 2026DIVERSED: Relaxed Speculative Decoding via Dynamic Ensemble Verificationcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Faster but Different: Diagnosing and Controlling Content Drift in Accelerated Multimodal Diffusion Language Modelscited by 0
  2. 2026Affix Cache for Diffusion Large Language Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Early-Bird Decoding: Accelerating Diffusion LLMs with Learnable Block Sizes and Parallel Sampling
  2. 2025Free Draft-and-Verification: Toward Lossless Parallel Decoding for Diffusion Large Language Models
  3. 2026Roll Out and Roll Back: Diffusion LLMs are Their Own Efficiency Teachers
  4. 2026CAI-DLLM: Convergence Aware Inference for Diffusion Language Models
  5. 2025Improving the Throughput of Diffusion-based Large Language Models via a Training-Free Confidence-Aware Calibration
  6. 2025From Bits to Rounds: Parallel Decoding with Exploration for Diffusion Language Models
  7. 2026DAWN: Dependency-Aware Fast Inference for Diffusion LLMs
  8. 2025Accelerating Diffusion LLM Inference via Local Determinism Propagation