EnSToM: Enhancing Dialogue Systems with Entropy-Scaled Steering Vectors for Topic Maintenance

TL;DR

EnSToM introduces entropy-scaled steering vectors to improve topic maintenance in small language models, effectively handling off-topic inputs while maintaining efficiency and accuracy in dialogue systems.

Contribution

The paper proposes a novel entropy-scaled steering vector method (EnSToM) that enhances topic adherence in small language models without extensive fine-tuning.

Findings

EnSToM significantly improves topic adherence in sLLMs.

EnSToM outperforms existing activation engineering methods.

The approach maintains efficiency and requires less data than fine-tuning.

Abstract

Small large language models (sLLMs) offer the advantage of being lightweight and efficient, which makes them suitable for resource-constrained environments. However, sLLMs often struggle to maintain topic consistency in task-oriented dialogue systems, which is critical for scenarios such as service chatbots. Specifically, it is important to ensure that the model denies off-topic or malicious inputs and adheres to its intended functionality so as to prevent potential misuse and uphold reliability. Towards this, existing activation engineering approaches have been proposed to manipulate internal activations during inference. While these methods are effective in certain scenarios, our preliminary experiments reveal their limitations in ensuring topic adherence. Therefore, to address this, we propose a novel approach termed Entropy-scaled Steering vectors for Topic Maintenance (EnSToM). EnSToM dynamically adjusts the steering intensity based on input uncertainty, which allows the model to handle off-topic distractors effectively while preserving on-topic accuracy. Our experiments demonstrate that EnSToM achieves significant performance gain with a relatively small data size compared to fine-tuning approaches. By improving topic adherence without compromising efficiency, our approach provides a robust solution for enhancing sLLM-based dialogue systems.