Controlling Large Language Model Agents with Entropic Activation Steering
This work addresses the challenge of controlling decision-making in LLM agents for AI and machine learning applications, representing an incremental advancement in activation steering methods.
The paper tackled the problem of controlling exploratory behaviors in large language model (LLM) agents by introducing Entropic Activation Steering (EAST), a method that manipulates high-level actions and modulates uncertainty, showing that LLM agents encode action uncertainty in their representation space.
The rise of large language models (LLMs) has prompted increasing interest in their use as in-context learning agents. At the core of agentic behavior is the capacity for exploration, or the ability to actively gather information about the environment. But how do LLM agents explore, and how can we control their exploratory behaviors? To answer these questions, we take a representation-level perspective, and introduce Entropic Activation Steering (EAST), an activation steering method for in-context LLM agents. Firstly, we demonstrate that EAST can effectively manipulate an LLM agent's exploration by directly affecting the high-level actions parsed from the outputs of the LLM, in contrast to token-level temperature sampling. Secondly, we reveal how applying this control modulates the uncertainty exhibited in the LLM's thoughts, guiding the agent towards more exploratory actions. Finally, we demonstrate that the steering vectors obtained by EAST generalize across task variants. In total, these results show that LLM agents explicitly encode uncertainty over their actions in their representation space. Our work paves the way for a new understanding of the functioning of LLM agents and to effective control of their decision-making behaviors.