mlsanityData to 14 Oct 2026

LOCKR: A Hidden-State Trajectory-Guided Planner for Detecting and Repairing Stable-but-Wrong Lock-In in Diffusion Language Models

arXiv 2609.27220 ↗First posted 23 Sep 2026cs.CLcited by 0

On the map: Diffusion models → Discrete and text diffusion

Abstract (arXiv)

Diffusion language models generate text through iterative denoising, exposing intermediate trajectories before final answers are produced. We identify a recurring reasoning failure, stable-but-wrong lock-in, where an answer stabilizes early around an incorrect value while substantial denoising remains. Surface-level decoding signals such as confidence, entropy, margin, and answer stability are insufficient to reliably distinguish correct from erroneous lock-in. We formulate selective reasoning repair as a lightweight test-time planning problem and propose LOCKR, a hidden-state trajectory-guided planner that decides when to allocate additional computation, expands a structured set of targeted repair branches, and selects the most promising continuation using trajectory-aware verification. Across two diffusion language models and three mathematical reasoning benchmarks, hidden-state trajectories consistently outperform surface signals and single hidden snapshots for both wrong-lock-in detection and repair selection. On natural evaluation distributions, LOCKR yields absolute accuracy gains of 2.21--5.37 percentage points across all five evaluated settings, with repair rates ranging from 22% to 41%. These results establish hidden diffusion trajectories as actionable signals for selective test-time reasoning repair.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Discrete and text diffusion, one of 14 directions in Diffusion models. Builds on 42 of its 56 references that are in the map. First posted on arXiv 23 Sep 2026.

Builds on

42 of 56 references.

Closest ancestors · 14

Same direction, newest first

  1. 2026Prism: Efficient Test-Time Scaling via Hierarchical Search and Self-Verification for Discrete Diffusion Language Models2 Feb
  2. 2025LLaDA2.0: Scaling Up Diffusion Language Models to 100B10 Dec
  3. 2025Fast-Decoding Diffusion Language Models via Progress-Aware Confidence Schedules2 Dec
  4. 2025Fine-Tuning Masked Diffusion for Provable Self-Correction1 Oct
  5. 2025Don't Settle Too Early: Self-Reflective Remasking for Diffusion Language Models28 Sep
  6. 2025Dream 7B: Diffusion Large Language Models21 Aug
  7. 2025Time Is a Feature: Exploiting Temporal Dynamics in Diffusion Language Models12 Aug
  8. 2025Diffusion LLMs Can Do Faster-Than-AR Inference via Discrete Diffusion Forcing8 Aug
  9. 2025Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models12 Mar
  10. 2025Remasking Discrete Diffusion Models with Inference-Time Scaling1 Mar
  11. 2025Large Language Diffusion Models14 Feb
  12. 2024Scaling Diffusion Language Models via Adaptation from Autoregressive Models23 Oct
2 more
  1. 2024Simple and Effective Masked Diffusion Language Models11 Jun
  2. 2022SSD-LM: Semi-autoregressive Simplex-based Diffusion Language Model for Text Generation and Modular Control31 Oct

Other references · 28

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  4. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
  5. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  6. 2017On Calibration of Modern Neural Networkscited by 4,960
  7. 2014Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translationcited by 4,057
  8. 2023Let's Verify Step by Stepcited by 3,789
  9. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,442
  10. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,211
  11. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameterscited by 1,986
  12. 2022Language Models (Mostly) Know What They Knowcited by 1,856
16 more
  1. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,719
  2. 2024Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvementcited by 1,090
  3. 2022Diffusion-LM Improves Controllable Text Generationcited by 1,027
  4. 2023Graph of Thoughts: Solving Elaborate Problems with Large Language Modelscited by 886
  5. 2023SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Modelscited by 810
  6. 2023MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Modelscited by 735
  7. 2020How Can We Know When Language Models Know? On the Calibration of Language Models for Question Answeringcited by 445
  8. 2025Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decodingcited by 420
  9. 2025The Lessons of Developing Process Reward Models in Mathematical Reasoningcited by 415
  10. 2024INSIDE: LLMs' Internal States Retain the Power of Hallucination Detectioncited by 323
  11. 2024ProcessBench: Identifying Process Errors in Mathematical Reasoningcited by 292
  12. 2025dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Cachingcited by 169
  13. 2024Orca-Math: Unlocking the potential of SLMs in Grade School Mathcited by 151
  14. 2024In-Context Sharpness as Alerts: An Inner Representation Perspective for Hallucination Mitigationcited by 45
  15. 2025ICR Probe: Tracking Hidden State Dynamics for Reliable Hallucination Detection in LLMscited by 32
  16. 2026HIVE: Hidden-Evidence Verification for Hallucination Detection in Diffusion Large Language Modelscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Beyond Mode-Seeking RL: Trajectory-Balance Post-Training for Diffusion Language Models
  2. 2025Reasoning in Diffusion Large Language Models is Concentrated in Dynamic Confusion Zones
  3. 2026Just on Time: Token-Level Early Stopping for Diffusion Language Models
  4. 2025Advancing Reasoning in Diffusion Language Models with Denoising Process Rewards
  5. 2025EDIT: Early Diffusion Inference Termination for dLLMs Based on Dynamics of Training Gradients
  6. 2026CAI-DLLM: Convergence Aware Inference for Diffusion Language Models
  7. 2026Remask, Don't Replace: Token-to-Mask Refinement in Diffusion Language Models
  8. 2026Think First, Diffuse Fast: Improving Diffusion Language Model Reasoning via Autoregressive Plan Conditioning