Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2601.09269 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911383722917888 |
|---|---|
| author | Ye, Wencheng Yuan, Xiaoyang Bin, Yi Zeng, Pengpeng Jin, Hengyu Peng, Liang Shen, Heng Tao |
| author_facet | Ye, Wencheng Yuan, Xiaoyang Bin, Yi Zeng, Pengpeng Jin, Hengyu Peng, Liang Shen, Heng Tao |
| contents | Recent work on domain-specific reasoning with large language models (LLMs) often relies on training-intensive approaches that require parameter updates. While activation steering has emerged as a parameter efficient alternative, existing methods apply static, manual interventions that fail to adapt to the dynamic nature of complex reasoning. To address this limitation, we propose RISER (Router-based Intervention for Steerable Enhancement of Reasoning), a plug-and-play intervention framework that adaptively steers LLM reasoning in activation space. RISER constructs a library of reusable reasoning vectors and employs a lightweight Router to dynamically compose them for each input. The Router is optimized via reinforcement learning under task-level rewards, activating latent cognitive primitives in an emergent and compositional manner. Across seven diverse benchmarks, RISER yields 3.4-6.5% average zero-shot accuracy improvements over the base model while surpassing CoT-style reasoning with 2-3x higher token efficiency and robust accuracy gains. Further analysis shows that RISER autonomously combines multiple vectors into interpretable, precise control strategies, pointing toward more controllable and efficient LLM reasoning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_09269 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | RISER: Orchestrating Latent Reasoning Skills for Adaptive Activation Steering Ye, Wencheng Yuan, Xiaoyang Bin, Yi Zeng, Pengpeng Jin, Hengyu Peng, Liang Shen, Heng Tao Artificial Intelligence Recent work on domain-specific reasoning with large language models (LLMs) often relies on training-intensive approaches that require parameter updates. While activation steering has emerged as a parameter efficient alternative, existing methods apply static, manual interventions that fail to adapt to the dynamic nature of complex reasoning. To address this limitation, we propose RISER (Router-based Intervention for Steerable Enhancement of Reasoning), a plug-and-play intervention framework that adaptively steers LLM reasoning in activation space. RISER constructs a library of reusable reasoning vectors and employs a lightweight Router to dynamically compose them for each input. The Router is optimized via reinforcement learning under task-level rewards, activating latent cognitive primitives in an emergent and compositional manner. Across seven diverse benchmarks, RISER yields 3.4-6.5% average zero-shot accuracy improvements over the base model while surpassing CoT-style reasoning with 2-3x higher token efficiency and robust accuracy gains. Further analysis shows that RISER autonomously combines multiple vectors into interpretable, precise control strategies, pointing toward more controllable and efficient LLM reasoning. |
| title | RISER: Orchestrating Latent Reasoning Skills for Adaptive Activation Steering |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2601.09269 |