mlsanityData to 14 Oct 2026

Plan Canvas: Fixed Reasoning Regions for Continuous Language Flows

arXiv 2610.05815 ↗First posted 5 Oct 2026cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

Continuous language flows generate text by denoising all positions of a target canvas together. The natural way to add reasoning to such a model is to write a trace ahead of the answer, but the trace length changes from question to question. The answer start is therefore unknown during denoising, and the model has to decide the trace length, the place of every trace token, and the answer at the same time. We propose Plan Canvas to fix the boundary between the trace and the answer. A plan region of fixed capacity holds a compact trace, supervised padding fills its unused positions, and the answer starts at a fixed position. The fixed regions also allow separate denoising clocks for the plan and for the answer. With the trace text, backbone, and canvas length of the free-trace baseline held fixed, Plan Canvas improves accuracy on ProsQA and on Deep ProsQA, a graph benchmark with longer proofs. On Deep ProsQA, accuracy rises from 73.0% to 87.0%, the share of questions answered with a valid path rises from 30.8% to 59.1%, and the gain is largest on the longest proofs.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 43 of its 44 references that are in the map. First posted on arXiv 5 Oct 2026.

Builds on

43 of 44 references.

References · 43

  1. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2022Classifier-Free Diffusion Guidancecited by 5,847
  4. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  5. 2022Flow Matching for Generative Modelingcited by 5,702
  6. 2022Large Language Models are Zero-Shot Reasonerscited by 4,374
  7. 2022Least-to-Most Prompting Enables Complex Reasoning in Large Language Modelscited by 1,407
  8. 2022Diffusion-LM Improves Controllable Text Generationcited by 1,027
  9. 2024Training Large Language Models to Reason in a Continuous Latent Spacecited by 797
  10. 2024Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusioncited by 681
  11. 2023Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Modelscited by 560
  12. 2022DiffuSeq: Sequence to Sequence Text Generation with Diffusion Modelscited by 450
31 more
  1. 2025Muon is Scalable for LLM Trainingcited by 428
  2. 2025Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approachcited by 415
  3. 2025Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Modelscited by 409
  4. 2022Analog Bits: Generating Discrete Data using Diffusion Models with Self-Conditioningcited by 373
  5. 2025Chain of Draft: Thinking Faster by Writing Lesscited by 297
  6. 2025CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillationcited by 225
  7. 2022Continuous diffusion for categorical datacited by 214
  8. 2024Chain-of-Thought Reasoning Without Promptingcited by 210
  9. 2024Compressed Chain of Thought: Efficient Reasoning Through Dense Representationscited by 201
  10. 2022SSD-LM: Semi-autoregressive Simplex-based Diffusion Language Model for Text Generation and Modular Controlcited by 194
  11. 2024Let's Think Dot by Dot: Hidden Computation in Transformer Language Modelscited by 191
  12. 2025Remasking Discrete Diffusion Models with Inference-Time Scalingcited by 189
  13. 2023Implicit Chain of Thought Reasoning via Knowledge Distillationcited by 176
  14. 2022Self-conditioned Embedding Diffusion for Text Generationcited by 117
  15. 2025Soft Thinking: Unlocking the Reasoning Potential of LLMs in Continuous Concept Spacecited by 117
  16. 2025SoftCoT: Soft Chain-of-Thought for Efficient Reasoning with LLMscited by 116
  17. 2024Self-Discover: Large Language Models Self-Compose Reasoning Structurescited by 110
  18. 2023TESS: Text-to-Text Self-Conditioned Simplex Diffusioncited by 65
  19. 2024Think While You Generate: Discrete Diffusion with Planned Denoisingcited by 63
  20. 2025Any-Order Flexible Length Masked Diffusioncited by 51
  21. 2026ELF: Embedded Language Flowscited by 45
  22. 2025Coevolutionary Continuous Discrete Diffusion: Make Your Diffusion Language Model a Latent Reasonercited by 32
  23. 2025LaDiR: Latent Diffusion Enhances LLMs for Text Reasoningcited by 29
  24. 2025Parallel Continuous Chain-of-Thought with Jacobi Iterationcited by 25
  25. 2026DreamOn: Diffusion Language Models For Code Infilling Beyond Fixed-size Canvascited by 19
  26. 2025ReFusion: A Diffusion Large Language Model with Parallel Autoregressive Decodingcited by 18
  27. 2026Parallelism and Generation Order in Masked Diffusion Language Models: Limits Today, Potential Tomorrowcited by 11
  28. 2025Planned Diffusioncited by 9
  29. 2026Latent Chain-of-Thought as Planning: Decoupling Reasoning from Verbalizationcited by 6
  30. 2026Answer First, Reason Later: When Commitment Order Costs Accuracy in Diffusion Language Modelscited by 3
  31. 2026From Interface to Inference: Eliciting Any-Order Inference from Any-Order Modelscited by 1

Built on by

Not cited yet.