EMNLP Findings2025
Can We Steer Reasoning Direction by Thinking Intervention?
Xingsheng Zhang, Luxi Xing, Chen Zhang, Yanbing Liu, Yifan Deng, Yunpeng Li, Yue Hu, Chenxu Niu
Abstract
Large Reason Models (LRMs) extend long reasoning process to solve complex tasks. However, due to the lack of fine-grained control, they often suffer from overthinking and erroneous reasoning problems, risking accuracy loss. To address this issue, we introduce Reasoning Direction Steering (RDS) to enable finegrained control over LRMs' reasoning behaviors by aligning reasoning trajectories with specific cognitive patterns. We develop a simple yet effective paradigm, Thinking Intervention, which explores two key dimensions -intervention positions and intervention stylesto achieve integration intervention throughout model reasoning processes. To validate the effectiveness of our approach, we conduct comprehensive experiments on multi-hop question answering tasks using state-of-the-art LRMs, including Qwen3-Series and R1-Series models. Experimental results demonstrate the efficacy of Thinking Intervention with 9.4% average improvement on R1-Series models and 1.9% improvement on Qwen3-Series models.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
1 top-tier paper has cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- LoRA: Low-Rank Adaptation of Large Language Models cites Language Models are Few-Shot Learners
- Locating and Editing Factual Associations in GPT cites Language Models are Few-Shot Learners
- Dynamic Early Exit in Reasoning Models cites Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
Cited by top-tier papers1
Builds on19
- Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah et al.NeurIPS 2020 · 64,255 citations
- Chain-of-Thought Prompting Elicits Reasoning in Large Language ModelsJason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma et al.NeurIPS 2022 · 22,562 citations
- LoRA: Low-Rank Adaptation of Large Language ModelsEdward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu et al.ICLR 2022 · 18,833 citations
- Locating and Editing Factual Associations in GPTKevin Meng, David Bau, Alex Andonian, Yonatan BelinkovNeurIPS 2022 · 3,415 citations
- Dynamic Early Exit in Reasoning ModelsChenxu Yang, Qingyi Si, Yongjie Duan, Zheliang Zhu et al.ICLR 2026 · 250 citations
Related papers
- Measuring the Faithfulness of Thinking Drafts in Large Reasoning ModelsZidi Xiong, Shan Chen, Zhenting Qi, Himabindu LakkarajuNeurIPS 2025 · 23 citations
- MuTIS: Enhancing Reasoning Efficiency through Multi Turn Intervention Sampling in Reinforcement LearningWenshuo Zhao, Haoxing Zhai, Xinyu Qiu, Zhenting Qi et al.EMNLP 2025 · 3 citations
- Intervene When It Doubts: Conjunction-Guided Interactive ReasoningQianyue Wang, Jinwu Hu, Yaofo Chen, Yufeng Wang et al.ICML 2026
- ThinkSwitcher: When to Think Hard, When to Think FastGuosheng Liang, Longguang Zhong, Ziyi Yang, Xiaojun QuanEMNLP Findings 2025 · 1 citation
- Modeling Hierarchical Thinking in Large Reasoning ModelsG M Shahariar, Erfan Shayegani, Ali Nazari, Nael Abu-GhazalehICML 2026
