CogSteer: Cognition-Inspired Selective Layer Intervention for Efficiently Steering Large Language Models
This work addresses the challenge of efficiently and safely deploying LLMs for specific applications, though it is incremental as it builds on existing parameter-efficient fine-tuning methods.
The authors tackled the problem of steering large language models (LLMs) more efficiently and interpretably by proposing a cognition-inspired method to select optimal layers for intervention, which improved performance on tasks like natural language understanding and generation across models like GPT-2 and Llama2-7B.
Large Language Models (LLMs) achieve remarkable performance through pretraining on extensive data. This enables efficient adaptation to diverse downstream tasks. However, the lack of interpretability in their underlying mechanisms limits the ability to effectively steer LLMs for specific applications. In this work, we investigate the intrinsic mechanisms of LLMs from a cognitive perspective using eye movement measures. Specifically, we analyze the layer-wise correlation between human cognitive indicators and LLM representations. Building on these insights, we propose a heuristic approach for selecting the optimal steering layer to modulate LLM semantics. To this end, we introduce an efficient selective layer intervention based on prominent parameter-efficient fine-tuning methods, which conventionally adjust either all layers or only the final layer. Additionally, we present an implicit layer contrastive intervention during inference to steer LLMs away from toxic outputs. Extensive experiments on natural language understanding, reasoning, and generation tasks, conducted on GPT-2, Llama2-7B, and Mistral-7B, demonstrate the effectiveness and efficiency of our approach. As a model-agnostic framework, it enhances the interpretability of LLMs while improving efficiency for safe deployment.