Understanding Reasoning in Thinking Language Models via Steering Vectors

MATS Fellow:

Constantin Venhoff

Authors:

Constantin Venhoff, Iván Arcuschin, Philip Torr, Arthur Conmy, Neel Nanda

Citations

66 Citations

Abstract:

Recent advances in large language models (LLMs) have led to the development of thinking language models that generate extensive internal reasoning chains before producing responses. While these models achieve improved performance, controlling their reasoning processes remains challenging. This work presents a steering approach for thinking LLMs by analyzing and manipulating specific reasoning behaviors in DeepSeek-R1-Distill models. Through a systematic experiment on 500 tasks across 10 diverse categories, we identify several reasoning behaviors exhibited by thinking models, including expressing uncertainty, generating examples for hypothesis validation, and backtracking in reasoning chains. We demonstrate that these behaviors are mediated by linear directions in the model's activation space and can be controlled using steering vectors. By extracting and applying these vectors, we provide a method to modulate specific aspects of the model's reasoning process, such as its tendency to backtrack or express uncertainty. Our approach offers practical tools for steering reasoning processes in thinking models in a controlled and interpretable manner. We validate our steering method using three DeepSeek-R1-Distill models, demonstrating consistent control across different model architectures.

Recent research

Subliminal Learning Is Steering Vector Distillation

Authors:

Date:

May 31, 2026

Citations:

Training Deliberative Monitors for Black-Box Scheming Detection

Authors:

Akshat Naik

Date:

May 28, 2026

Citations:

Frequently asked questions

What is the MATS Program?
How long does the program last?