# Agentic Chain-of-Thought Steering for Efficient and Controllable LLM Reasoning

arXiv: https://arxiv.org/abs/2606.03965 · HTML: https://mlsanity.com/paper/2606.03965/

## Key facts

- Cited by 1 papers 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 21 of its 57 references that are in the map.
- First posted on arXiv 2 Jun 2026.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)
- LLM agents and tool use / [General LLM agent frameworks and benchmarks](https://mlsanity.com/agents/general-llm-agent-frameworks-and-benchmarks/)

## Builds on (most cited first, 21 of 21)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Understanding R1-Zero-Like Training: A Critical Perspective](https://mlsanity.com/paper/2503.20783/) (2025)
- [Chain of Draft: Thinking Faster by Writing Less](https://mlsanity.com/paper/2502.18600/) (2025)
- [O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning](https://mlsanity.com/paper/2501.12570/) (2025)
- [Reasoning Models Can Be Effective Without Thinking](https://mlsanity.com/paper/2504.09858/) (2025)
- [Llama-Nemotron: Efficient Reasoning Models](https://mlsanity.com/paper/2505.00949/) (2025)
- [Concise Thoughts: Impact of Output Length on LLM Reasoning and Cost](https://mlsanity.com/paper/2407.19825/) (2024)
- [Towards System 2 Reasoning in LLMs: Learning How to Think With Meta Chain-of-Thought](https://mlsanity.com/paper/2501.04682/) (2025)
- [OpenRLHF: An Easy-to-use, Scalable and High-performance RLHF Framework](https://mlsanity.com/paper/2405.11143/) (2024)
- [Understanding Reasoning in Thinking Language Models via Steering Vectors](https://mlsanity.com/paper/2506.18167/) (2025)
- [Meta-Reasoner: Dynamic Guidance for Optimized Inference-time Reasoning in Large Language Models](https://mlsanity.com/paper/2502.19918/) (2025)
- [Steering LLM Thinking with Budget Guidance](https://mlsanity.com/paper/2506.13752/) (2025)
- [BudgetThinker: Empowering Budget-aware LLM Reasoning with Control Tokens](https://mlsanity.com/paper/2508.17196/) (2025)
- [FlashThink: An Early Exit Method For Efficient Reasoning](https://mlsanity.com/paper/2505.13949/) (2025)
- [RL of Thoughts: Navigating LLM Reasoning with Inference-time Reinforcement Learning](https://mlsanity.com/paper/2505.14140/) (2025)
- [Learning to Hint for Reinforcement Learning](https://mlsanity.com/paper/2604.00698/) (2026)
- [Understanding and Steering the Cognitive Behaviors of Reasoning Models at Test-Time](https://mlsanity.com/paper/2512.24574/) (2025)
- [Neural Chain-of-Thought Search: Searching the Optimal Reasoning Path to Enhance Large Language Models](https://mlsanity.com/paper/2601.11340/) (2026)
- [ThinkRouter: Efficient Reasoning via Routing Thinking between Latent and Discrete Spaces](https://mlsanity.com/paper/2602.11683/) (2026)
- [SmartThinker: Learning to Compress and Preserve Reasoning by Step-Level Length Control](https://mlsanity.com/paper/2507.04348/) (2025)

## Built on by (most cited first, 1 of 1)

- [MetaCtrl: Your Large Language Models Can Reason Better and More Concisely with a Metacognitive Controller](https://mlsanity.com/paper/2609.37304/) (2026)

## Abstract (arXiv)

Large language models improve final-answer accuracy through extended chain-of-thought reasoning, but often spend tokens inefficiently and offer little inference-time control. Existing efficient reasoning methods control thinking length by shortening, early-stopping, or compressing traces, leaving how the model thinks implicit. In this paper, we propose Agentic Chain-of-Thought Steering (ACTS), which formulates reasoning steering as a Markov decision process where a controller agent adaptively steers a frozen reasoner during inference. At each step, the controller observes the reasoning trace and remaining thinking budget, then issues a steering action consisting of a reasoning strategy and a steering phrase that initiates the next reasoner step. This enables budget-aware strategy control for efficient reasoning while preserving the reasoner's generation continuity. We initialize the controller agent from our constructed synthetic steering trajectories with multi-budget augmentation, and further optimize it via reinforcement learning with budget-conditioned reward shaping. Experiments across multiple benchmarks show that ACTS achieves competitive accuracy with substantial token savings, and enables controllable accuracy-token trade-offs across different reasoners and tasks. The code is available at https://github.com/Andree-9/ACTS.