mlsanityData to 14 Oct 2026

Search for Truth from Reasoning: A Dynamic Representation Editing Framework for Steering LLM Trajectories

arXiv 2606.28589 ↗First posted 26 Jun 2026cs.AIcited by 1

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

Current approaches to enhance Large Language Model (LLM) reasoning, such as Chain-of-Thought and "Wait" prompts, primarily encourage models to think more, yet often fail to guide them toward Truth. While Representation Editing (RepE) offers a intrinsic control, its application to dynamic reasoning trajectories remains underexplored. In this work, we bridge this gap by investigating the geometry of truth within unfolding reasoning chains. We uncover three critical insights: (1) Truth is encoded at the sentence level and is entangled with latent reasoning patterns; (2) Effective intervention follows an Uncertainty Principle and a Decay Effect, requiring localization to early, high-entropy forks; (3) Naive steering vectors suffer from noise, risking collateral damage to correct trajectories. Based on these findings, we propose DynaSteer, a dynamic RepE framework. DynaSteer employs pattern clustering to disentangle reasoning manifolds and utilizes Fisher-LDA to project purified truth. By dynamically monitoring lookahead entropy, it selectively steers and rolls back trajectories only when necessary. Comprehensive experimental results on several MATH benchmark verify the effectiveness of DynaSteer, and experiments on out-of-domain coding tasks further confirm its generalization ability. Our code is publicly available at https://github.com/tianlwang/DynaSteer.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 24 of its 48 references that are in the map. First posted on arXiv 26 Jun 2026.

Builds on

24 of 48 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025Wait, Wait, Wait... Why Do Reasoning Models Loop?15 Dec
  2. 2025Understanding Aha Moments: from External Observations to Internal Mechanisms3 Apr

Other references · 22

  1. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  2. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  3. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  4. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  5. 2021Program Synthesis with Large Language Modelscited by 3,637
  6. 2023Qwen Technical Reportcited by 3,344
  7. 2019HuggingFace's Transformers: State-of-the-art Natural Language Processingcited by 2,218
  8. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,336
  9. 2023The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasetscited by 712
  10. 2025Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoningcited by 559
  11. 2024GSM-Plus: A Comprehensive Benchmark for Evaluating the Robustness of LLMs as Mathematical Problem Solverscited by 158
  12. 2024Multi-Step Reasoning with Large Language Models, a Surveycited by 143
10 more
  1. 2025Thought Anchors: Which LLM Reasoning Steps Matter?cited by 129
  2. 2025Rethinking Entropy Interventions in RLVR: An Entropy Change Perspectivecited by 52
  3. 2024Entropy Law: The Story Behind Data Compression and LLM Performancecited by 44
  4. 2025ProMed: Shapley Information Gain Guided Reinforcement Learning for Proactive Medical LLMscited by 13
  5. 2025DRESSing Up LLM: Efficient Stylized Question-Answering via Style Subspace Editingcited by 10
  6. 2025Detoxifying Large Language Models via Autoregressive Reward Guided Representation Editingcited by 10
  7. 2025LearNAT: Learning NL2SQL with AST-guided Task Decomposition for Large Language Modelscited by 8
  8. 2025ADEPT: Continual Pretraining via Adaptive Expansion and Dynamic Decoupled Tuningcited by 5
  9. 2025Balancing Stylization and Truth via Disentangled Representation Steeringcited by 2
  10. 2026GraphWalker: Patient Analogy Meets Information Gain for Clinical Reasoning with Large Language Modelscited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026The Imitation Game: When LLMs Learn to Reason Like Programs via Code-Centric Reasoning Data Synthesiscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Truth as a Trajectory: What Internal Representations Reveal About Large Language Model Reasoning
  2. 2026LLM Reasoning as Trajectories: Step-Specific Representation Geometry and Correctness Signals
  3. 2026Invariant Reasoning Directions in Latent Trajectories of Language Models
  4. 2026Can We Break LLMs Out of Self-Loops? Fine-Grained Reasoning Control with Activation Steering
  5. 2026Reasoning emerges from constrained inference manifolds in large language models
  6. 2026Answer-Distribution Trajectories: A Stochastic-Dynamics View of LLM Reasoning
  7. 2025From Implicit Exploration to Structured Reasoning: Leveraging Guideline and Refinement for LLMs
  8. 2026Confidence Geometry Reveals Trace-Level Correctness in Large Language Model Reasoning