paper-with-me

홈 › Papers

Reflection Steering: Disentangling Reflection from Reasoning in Activation Space for Token-Efficient Inference

2026-08-26 · Jiarui Hu, Zhiyuan Wen, Xiaoyun Liu, Jiaxing Shen, Yu Yang arxiv

Large reasoning models often produce reasoning traces with verification, revision, and backtracking. When reflection merely re-checks established results, it wastes reasoning tokens and increases latency. Most existing reflection steering methods add a label-derived mean-difference direction across preset layers, but its entanglement with reasoning and length signals destabilizes the accuracy-efficiency trade-off. In this paper, we propose Reflection Steering, a training-free framework for controlling reflection-associated computation within LLMs by disentangling reflection-related activations from general reasoning. Specifically, we contrast reflective and non-reflective hidden states at each LLM layer, denoise the resulting reflection directions with PCA, and orthogonalize them against general-reasoning directions. To limit downstream amplification from early-layer interventions, we calibrate each layer across multiple intervention strengths on a small set, retain only stable layers, and apply bounded projection removal to their residual-stream activations. We conduct extensive experiments across two public benchmarks and three open-weight LLMs against state-of-the-art activation-steering baselines. Results show that Reflection Steering reduces reasoning tokens by 16.9% on average across six matched settings. Besides, our method further introduces a bounded reflection intervention-strength parameter $α$, enabling deployment-time adjustment to balance token savings, accuracy, and generation stability.

📄 PDF Abstract BibTeX arXiv:2608.25542

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Unveiling the Latent Directions of Reflection in Large Language Models

2025-08-23 · Fu-Chieh Chang, Yu-Ting Lee, Pei-Yuan Wu arxiv

Reflection, the ability of large language models (LLMs) to evaluate and revise their own reasoning, has been widely used to improve performance on complex reasoning tasks. Yet, most prior works emphasizes designing refle…

Reinforcement Learning

ReflCtrl: Controlling LLM Reflection via Representation Engineering

2025-12-16 · Ge Yan, Chung-En Sun, Tsui-Wei, Weng arxiv

Large language models (LLMs) with Chain-of-Thought (CoT) reasoning have achieved strong performance across diverse tasks, including mathematics, coding, and general reasoning. A distinctive ability of these reasoning mod…

From Emergence to Control: Probing and Modulating Self-Reflection in Language Models

2025-06-13 · XUDONG ZHU, Jiachen Jiang, Mohammad Mahdi Khalili, Zhihui Zhu

Self-reflection -- the ability of a large language model (LLM) to revisit, evaluate, and revise its own reasoning -- has recently emerged as a powerful behavior enabled by reinforcement learning with verifiable rewards (…

Large Language ModelNavigate

Fantastic Reasoning Behaviors and Where to Find Them: Unsupervised Discovery of the Reasoning Process

2025-12-30 · Zhenyu Zhang, Shujian Zhang, John Lambert, Wenxuan Zhou 외 arxiv

Despite the growing reasoning capabilities of recent large language models (LLMs), their internal mechanisms during the reasoning process remain underexplored. Prior approaches often rely on human-defined concepts (e.g.,…

Self-Guard: Defending Large Reasoning Models via enhanced self-reflection

2026-01-31 · Jingnan Zheng, Jingjun Xu, Yanzhen Luo, Chenhang Cui 외 arxiv

The emergence of Large Reasoning Models (LRMs) introduces a new paradigm of explicit reasoning, enabling remarkable advances yet posing unique risks such as reasoning manipulation and information leakage. To mitigate the…