mlsanityData to 7 Oct 2026

Break the Block: Dynamic-size Reasoning Blocks for Diffusion Large Language Models via Monotonic Entropy Descent with Reinforcement Learning

arXiv 2605.02263 ↗First posted 4 May 2026cs.LGcited by 1

On the map: LLM reasoning → Diffusion reasoning and test-time scaling

Abstract (arXiv)

Recent diffusion large language models (dLLMs) have demonstrated both effectiveness and efficiency in reasoning via a block-based semi-autoregressive generation paradigm. Despite their progress, the fixed-size block generations remain a critical bottleneck for effective and coherent reasoning. 1. From a global perspective, different reasoning tasks would correspond to different optimal decoding block sizes, which makes a ``one-size-fits-all'' assumption ineffective. 2. Even within a single reasoning task, the rigid block partitioning would break the logical flow and reduce reasoning coherence. Through empirical observations, we reveal that for block-wise entropy, incorrect reasoning exhibits a fluctuating and unsteady trend between blocks, whereas the correctly generated tasks follow a consistent descending trend. Therefore, this paper proposes b1, a novel post-training framework for dLLMs that learns dynamic-size reasoning blocks via a Monotonic Entropy Descent objective with reinforcement learning to enhance reasoning coherence.b1 integrates seamlessly as a plug-and-play module with existing dLLM's post-training algorithms. Extensive experiments across various reasoning benchmarks showcase b1's consistent improvement over existing fixed-size block baselines. Our code has been released at https://github.com/YanJiangJerry/Block-R1.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in Diffusion reasoning and test-time scaling, one of 17 directions in LLM reasoning. Builds on 25 of its 31 references that are in the map. First posted on arXiv 4 May 2026.

Builds on

25 of 31 references.

Closest ancestors · 4

Same direction, newest first

  1. 2025Improving Reasoning for Diffusion Language Models via Group Diffusion Policy Optimization9 Oct
  2. 2025d2: Improving Reasoning in Diffusion Language Models via Trajectory Likelihood Estimation25 Sep
  3. 2025wd1: Weighted Policy Optimization for Reasoning in Diffusion Language Models7 Jul
  4. 2025d1: Scaling Reasoning in Diffusion Large Language Models via Reinforcement Learning16 Apr

Other references · 21

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2017Proximal Policy Optimization Algorithmscited by 14,441
  3. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  4. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  5. 2023Let's Verify Step by Stepcited by 3,598
  6. 2023FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioningcited by 2,551
  7. 2025Large Language Diffusion Modelscited by 872
  8. 2025LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Modelscited by 273
  9. 2025MMaDA: Multimodal Large Diffusion Language Modelscited by 241
  10. 2025The Surprising Effectiveness of Negative Reinforcement in LLM Reasoningcited by 158
  11. 2025A Minimalist Approach to LLM Reasoning: from Rejection Sampling to Reinforcecited by 141
  12. 2025KodCode: A Diverse, Challenging, and Verifiable Synthetic Dataset for Codingcited by 115
9 more
  1. 2025Reinforcement Learning with Verifiable Rewards: GRPO's Effective Loss, Dynamics, and Success Amplificationcited by 108
  2. 2025Revolutionizing Reinforcement Learning Framework for Diffusion Large Language Modelscited by 80
  3. 2025NFT: Bridging Supervised Learning and Reinforcement Learning in Math Reasoningcited by 38
  4. 2025Fine-Tuning Discrete Diffusion Models with Policy Gradient Methodscited by 36
  5. 2025MDPO: Overcoming the Training-Inference Divide of Masked Diffusion Language Modelscited by 32
  6. 2025AdaBlock-dLLM: Semantic-Aware Diffusion LLM Inference via Adaptive Block Sizecited by 30
  7. 2026Multiplex Thinking: Reasoning via Token-wise Branch-and-Mergecited by 16
  8. 2025DiRL: An Efficient Post-Training Framework for Diffusion Language Modelscited by 10
  9. 2025Towards Principled Learning for Re-ranking in Recommender Systemscited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Block-R1: Rethinking the Role of Block Size in Multi-domain Reinforcement Learning for Diffusion Large Language Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Reasoning in Diffusion Large Language Models is Concentrated in Dynamic Confusion Zones
  2. 2025Enhancing Reasoning for Diffusion LLMs via Distribution Matching Policy Optimization
  3. 2026DreamReasoner-8B: Block-Size Curriculum Learning for Diffusion Reasoning Models
  4. 2025Training Language Models to Reason Efficiently
  5. 2025Beyond Surface Reasoning: Unveiling the True Long Chain-of-Thought Capacity of Diffusion Large Language Models
  6. 2025RFG: Test-Time Scaling for Diffusion Large Language Model Reasoning with Reward-Free Guidance
  7. 2025Advancing Reasoning in Diffusion Language Models with Denoising Process Rewards
  8. 2026Adaptive Test-Time Compute Allocation for Block Diffusion Language Models in Complex Reasoning