mlsanityData to 7 Oct 2026

Reliable Control-Point Selection for Steering Reasoning in Large Language Models

arXiv 2604.02113 ↗First posted 2 Apr 2026cs.CLcited by 1

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

Abstract (arXiv)

Steering vectors offer a training-free mechanism for controlling reasoning behaviors in large language models, but constructing effective vectors requires identifying genuine behavioral signals in the model's hidden states. For behaviors that can be toggled via prompts, this is straightforward. However, many reasoning behaviors -- such as self-reflection -- emerge spontaneously and resist prompt-level control. Current methods detect these behaviors through keyword matching in chain-of-thought traces, implicitly assuming that every detected boundary encodes a genuine behavioral signal. We show that this assumption is overwhelmingly wrong: across 541 keyword-detected boundaries, 93.3% are behaviorally unstable, failing to reproduce the detected behavior under re-generation from the same prefix. We develop a probabilistic model that formalizes intrinsic reasoning behaviors as stochastic events with context-dependent trigger probabilities, and show that unstable boundaries dilute the steering signal. Guided by this analysis, we propose stability filtering, which retains only boundaries where the model consistently reproduces the target behavior. Combined with a content-subspace projection that removes residual question-specific noise, our method achieves 0.784 accuracy on MATH-500 (+5.0 over the strongest baseline). The resulting steering vectors transfer across models in the same architecture family without re-extraction, improving Nemotron-Research-Reasoning-1.5B (+5.0) and DeepScaleR-1.5B-Preview (+6.0). Code is available at https://github.com/zhmzm/stability-steering.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 21 of its 22 references that are in the map. First posted on arXiv 2 Apr 2026.

Builds on

21 of 22 references.

Closest ancestors · 8

Same direction, newest first

  1. 2025Small Vectors, Big Effects: A Mechanistic Study of RL-Induced Reasoning via Steering Vectors8 Sep
  2. 2025Understanding Reasoning in Thinking Language Models via Steering Vectors22 Jun
  3. 2025Mitigating Overthinking in Large Reasoning Models via Manifold Steering28 May
  4. 2025SEAL: Steerable Reasoning Calibration of Large Language Models for Free7 Apr
  5. 2025Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models20 Mar
  6. 2025Revisiting the Test-Time Scaling of o1-like Models: Do they Truly Possess Test-Time Scaling Capabilities?17 Feb
  7. 2023Large Language Models Cannot Self-Correct Reasoning Yet3 Oct
  8. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Models17 May

Other references · 13

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
  2. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,601
  3. 2023Inference-Time Intervention: Eliciting Truthful Answers from a Language Modelcited by 1,259
  4. 2022Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 smallcited by 1,130
  5. 2023Steering Llama 2 via Contrastive Activation Additioncited by 1,009
  6. 2024Refusal in Language Models Is Mediated by a Single Directioncited by 960
  7. 2023Steering Language Models With Activation Engineeringcited by 824
  8. 2022Discovering Language Model Behaviors with Model-Written Evaluationscited by 805
  9. 2019Plug and Play Language Models: A Simple Approach to Controlled Text Generationcited by 793
  10. 2022Discovering Latent Knowledge in Language Models Without Supervisioncited by 733
  11. 2023The Linear Representation Hypothesis and the Geometry of Large Language Modelscited by 679
  12. 2020Dataset Cartography: Mapping and Diagnosing Datasets with Training Dynamicscited by 423
1 more
  1. 2025Persona Vectors: Monitoring and Controlling Character Traits in Language Modelscited by 303

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Adaptive Triggering for Bias Correction in LLM Reasoningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Predicting Future Behaviors in Reasoning Models Enables Better Steering
  2. 2026Can We Break LLMs Out of Self-Loops? Fine-Grained Reasoning Control with Activation Steering
  3. 2025Understanding and Steering the Cognitive Behaviors of Reasoning Models at Test-Time
  4. 2025Beyond Prompt Engineering: Robust Behavior Control in LLMs via Steering Target Atoms
  5. 2025KV Cache Steering for Controlling Frozen LLMs
  6. 2026A Coin Flip Per Token: Bernoulli Sparse Steering of Large Language Models
  7. 2026InjectRBP: Steering Large Language Model Reasoning Behavior via Pattern Injection
  8. 2026On the Generalization of Steering Vectors for Chain-of-Thought Faithfulness