mlsanityData to 14 Oct 2026

Generative Recursive Reasoning

arXiv 2605.19376 ↗First posted 19 May 2026cs.AIcited by 20

Not on the map. Why?

Abstract (arXiv)

How should future neural reasoning systems implement extended computation? Recursive Reasoning Models (RRMs) offer a promising alternative to autoregressive sequence extension by performing iterative latent-state refinement with shared transition functions. Yet existing RRMs are largely deterministic, following a single latent trajectory and converging to a single prediction. We introduce Generative Recursive reAsoning Models (GRAM), a framework that turns recursive latent reasoning into probabilistic multi-trajectory computation. GRAM models reasoning as a stochastic latent trajectory, enabling multiple hypotheses, alternative solution strategies, and inference-time scaling through both recursive depth and parallel trajectory sampling. This yields a latent-variable generative model supporting conditional reasoning via pθ(y mid x) and, with fixed or absent inputs, unconditional generation via pθ(x). Trained with amortized variational inference, GRAM improves over deterministic recurrent and recursive baselines on structured reasoning and multi-solution constraint satisfaction tasks, while demonstrating an unconditional generation capability. https://ahn-ml.github.io/gram-website

Key facts. Cited by 20 papers in the map (data through 14 Oct 2026). Builds on 47 of its 64 references that are in the map. First posted on arXiv 19 May 2026.

Builds on

47 of 64 references.

References · 47

  1. 2017Attention Is All You Needcited by 60,361
  2. 2017Proximal Policy Optimization Algorithmscited by 15,215
  3. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  4. 2022Scalable Diffusion Models with Transformerscited by 6,054
  5. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  6. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,703
  7. 2014Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translationcited by 4,057
  8. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,211
  9. 2016Improved Techniques for Training GANscited by 2,970
  10. 2019ALBERT: A Lite BERT for Self-supervised Learning of Language Representationscited by 2,790
  11. 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,626
  12. 2015High-Dimensional Continuous Control Using Generalized Advantage Estimationcited by 2,026
35 more
  1. 2020GLU Variants Improve Transformercited by 1,804
  2. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,719
  3. 2018Group Normalizationcited by 1,524
  4. 2018Learning Latent Dynamics for Planning from Pixelscited by 1,264
  5. 2023Mastering Diverse Domains through World Modelscited by 1,253
  6. 2020Improved Techniques for Training Score-Based Generative Modelscited by 1,043
  7. 2020Mastering Atari with Discrete World Modelscited by 1,007
  8. 2017Sigmoid-Weighted Linear Units for Neural Network Function Approximation in Reinforcement Learningcited by 950
  9. 2023Graph of Thoughts: Solving Elaborate Problems with Large Language Modelscited by 886
  10. 2024Simple and Effective Masked Diffusion Language Modelscited by 848
  11. 2024Training Large Language Models to Reason in a Continuous Latent Spacecited by 797
  12. 2016Adaptive Computation Time for Recurrent Neural Networkscited by 506
  13. 2025Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approachcited by 415
  14. 2015A Recurrent Latent Variable Model for Sequential Datacited by 362
  15. 2019Depth-Adaptive Transformercited by 236
  16. 2018A Note on the Inception Scorecited by 226
  17. 2025CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillationcited by 225
  18. 2025ARC-AGI-2: A New Challenge for Frontier AI Reasoning Systemscited by 190
  19. 2025Hierarchical Reasoning Modelcited by 177
  20. 2015Deep Kalman Filterscited by 166
  21. 2023Looped Transformers are Better at Learning Learning Algorithmscited by 149
  22. 2014A Tutorial on Principal Component Analysiscited by 141
  23. 2025Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computationcited by 133
  24. 2025Soft Thinking: Unlocking the Reasoning Potential of LLMs in Continuous Concept Spacecited by 117
  25. 2016Sequential Neural Models with Stochastic Layerscited by 115
  26. 2017The Consciousness Priorcited by 107
  27. 2024Relaxed Recursive Transformers: Effective Parameter Sharing with Layer-wise LoRAcited by 78
  28. 2025Reasoning by Superposition: A Theoretical Perspective on Chain of Continuous Thoughtcited by 77
  29. 2019Graph Colouring Meets Deep Learning: Effective Graph Neural Network Models for Combinatorial Problemscited by 40
  30. 2017Unbiasing Truncated Backpropagation Through Timecited by 37
  31. 2025Continuous Chain of Thought Enables Parallel Exploration and Reasoningcited by 35
  32. 2025Soft Tokens, Hard Truthscited by 35
  33. 2025Hybrid Latent Reasoning via Reinforcement Learningcited by 32
  34. 2023CoTFormer: A Chain-of-Thought Driven Architecture with Budget-Adaptive Computation Cost at Inferencecited by 27
  35. 2025Text Generation Beyond Discrete Token Samplingcited by 21

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 20.

20 papers.

Sort:

By area:

  1. 2026Bridging the Gap Between Latent and Explicit Reasoning with Looped Transformerscited by 13
  2. 2026Thinking with Looped Flowscited by 9
  3. 2026Retrofitting Recurrent Depth into a Pretrained Language Model: Installation, Extrapolation, Transfer, and Retention at Two Parameter Budgetscited by 6
  4. 2026Quantizing Recursive Reasoning Modelscited by 2
  5. 2026Diffusion as a Training Curriculum for Timestep-Free Iterative Reasoningcited by 2
  6. 2026Improving Test-Time Scaling with Adaptive Looped Transformerscited by 1
  7. 2026Grow-Prune-Freeze Networks: Adaptive & Continual Learning Technique for Olfactory Navigationcited by 0
  8. 2026DiARC: Distinguishing Positive and Negative Samples Helps Improving ARC-like Reasoning Ability of Large Language Modelscited by 0
  9. 2026Anatomy of a Sound Neural Reasoner: One-Shot Amortization, First-Pass Poisoning, and Search Inertness in Clue-Rich Completioncited by 0
  10. 2026Recursive Vision Language Models for General Symbolic Reasoningcited by 0
  11. 2026Latent Recurrent Thoughts: Recurrent Refinement of Proposed Latents for Reasoning with Frozen LLMscited by 0
  12. 2026Dissecting Hierarchical Reasoning Models: A Mechanistic Studycited by 0
8 more
  1. 2026Constraints Are Graphs, Not Chains: Exact Decoding for Diffusion Language Modelscited by 0
  2. 2026Reasoning with Neural Cellular Automatacited by 0
  3. 2026What Matters for Latent Reasoning with Flow Matchingcited by 0
  4. 2026Noise Your Prompt: Noising Conditioning Tokens in Continuous Diffusion Language Modelscited by 0
  5. 2026Unrolled Flow Models for Reasoningcited by 0
  6. 2026Invariant-Measure Reasoners: Stable Representations for Latent Reasoningcited by 0
  7. 2026Resample the Answer: Where to Spend Test-Time Compute in Recursive Reasoningcited by 0
  8. 2026Train for Many, Update with One: Improving Multi-Sample Reasoning via Fenchel-Bregman Posterior Learningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Boosting Inference with Guided Reasoning: Stochastic Exploration for Recursive Models
  2. 2026What Limits Recursive Reasoning Models: Optimization, Architecture and Test-Time Scaling
  3. 2026Dynamical Systems Theory Behind a Hierarchical Reasoning Model
  4. 2026Form Follows Function: Recursive Stem Model
  5. 2026Random Recursive Models
  6. 2026Energy-guided Recursive Model
  7. 2026Recursive Models for Long-Horizon Reasoning
  8. 2026Think Deep, Speak Once: Relit, A Recursive Latent Implicit Transformer Framework