mlsanityData to 7 Oct 2026

Swordsman: Entropy-Driven Adaptive Block Partition for Efficient Diffusion Language Models

arXiv 2602.04399 ↗First posted 4 Feb 2026cs.CLcited by 3

On the map: Diffusion models → Discrete and text diffusion

Abstract (arXiv)

Block-wise decoding effectively improves the inference speed and quality in diffusion language models (DLMs) by combining inter-block sequential denoising and intra-block parallel unmasking. However, existing block-wise decoding methods typically partition blocks in a rigid and fixed manner, which inevitably fragments complete semantic or syntactic constituents, leading to suboptimal performance. Inspired by the entropy reduction hypothesis (ERH), we recognize that constituent boundaries offer greater opportunities for uncertainty reduction, which motivates us to employ entropy analysis for identifying constituent boundaries. Therefore, we propose Swordsman, an entropy-driven adaptive block-wise decoding framework for DLMs. Swordsman adaptively partitions blocks by identifying entropy shifts between adjacent tokens to better align with semantic or syntactic constituent boundaries. In addition, Swordsman dynamically adjusts unmasking thresholds conditioned on the real-time unmasking status within a block, further improving both efficiency and stability. As a training-free framework, supported by KV Cache, Swordsman demonstrates state-of-the-art performance across extensive evaluations.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Placed in Discrete and text diffusion, one of 14 directions in Diffusion models. Builds on 30 of its 34 references that are in the map. First posted on arXiv 4 Feb 2026.

Builds on

30 of 34 references.

Closest ancestors · 18

Same direction, newest first

  1. 2026Parallelism and Generation Order in Masked Diffusion Language Models: Limits Today, Potential Tomorrow22 Jan
  2. 2025CDLM: Consistency Diffusion Language Models For Faster Sampling24 Nov
  3. 2025dInfer: An Efficient Inference Framework for Diffusion Language Models9 Oct
  4. 2025LLaDA-MoE: A Sparse MoE Diffusion Language Model29 Sep
  5. 2025Any-Order Flexible Length Masked Diffusion31 Aug
  6. 2025Dream 7B: Diffusion Large Language Models21 Aug
  7. 2025Empirical Analysis of Decoding Biases in Masked Diffusion Models18 Aug
  8. 2025A Survey on Diffusion Language Models14 Aug
  9. 2025Diffusion LLMs Can Do Faster-Than-AR Inference via Discrete Diffusion Forcing8 Aug
  10. 2025Accelerated Sampling from Masked Diffusion Models via Entropy Bounded Unmasking30 May
  11. 2025FlashDLM: Accelerating Diffusion Language Model Inference via Efficient KV Caching and Guided Diffusion27 May
  12. 2025Reinforcing the Diffusion Chain of Lateral Thought with Diffusion Language Models15 May
6 more
  1. 2025Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models12 Mar
  2. 2025Remasking Discrete Diffusion Models with Inference-Time Scaling1 Mar
  3. 2025Large Language Diffusion Models14 Feb
  4. 2024Simple and Effective Masked Diffusion Language Models11 Jun
  5. 2023Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution25 Oct
  6. 2022DiffuSeq: Sequence to Sequence Text Generation with Diffusion Models17 Oct

Other references · 12

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,665
  3. 2025Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decodingcited by 404
  4. 2025LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Modelscited by 273
  5. 2025Mercury: Ultra-Fast Language Models Based on Diffusioncited by 225
  6. 2025Fast-dLLM v2: Efficient Block-Diffusion LLMcited by 128
  7. 2025Accelerating Diffusion LLMs via Adaptive Parallel Decodingcited by 108
  8. 2025dParallel: Learnable Parallel Decoding for dLLMscited by 59
  9. 2025Wide-In, Narrow-Out: Revokable Decoding for Efficient and Effective DLLMscited by 45
  10. 2025AdaBlock-dLLM: Semantic-Aware Diffusion LLM Inference via Adaptive Block Sizecited by 30
  11. 2025Adaptive Classifier-Free Guidance via Dynamic Low-Confidence Maskingcited by 21
  12. 2025Structuring The Future: Diffusion LLM Speculative Decoding via Calibrated Draft Graphscited by 21

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 3.

3 papers.

  1. 2026DepCap: Adaptive Block-Wise Parallel Decoding for Efficient Diffusion LM Inferencecited by 1
  2. 2026Early-Bird Decoding: Accelerating Diffusion LLMs with Learnable Block Sizes and Parallel Samplingcited by 0
  3. 2026Unmask the State: When Does State Adaptation Matter for Masked Diffusion Language Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Adaptive Block Diffusion: Resolving Training-Inference Mismatch in Diffusion Language Models
  2. 2026S2D2: Fast Decoding for Diffusion LLMs via Training-Free Self-Speculation
  3. 2026SemBlock: Semantic Boundary Dynamic Blocks for Diffusion LLMs
  4. 2025Beyond Confidence: Adaptive and Coherent Decoding for Diffusion Language Models
  5. 2025From Bits to Rounds: Parallel Decoding with Exploration for Diffusion Language Models
  6. 2026Rethinking the Generation Order of Block Diffusion Language Models
  7. 2026Confidence-Based Decoding is Provably Efficient for Diffusion Language Models
  8. 2025Sequential Diffusion Language Models