Task-driven Layerwise Additive Activation Intervention

TL;DR

This paper introduces a layer-wise additive activation intervention framework that improves the efficiency and effectiveness of adapting language models to new tasks by optimizing activation manipulations.

Contribution

It presents a novel, optimized intervention method that reduces reliance on heuristics and prompts, enhancing task adaptation in language models.

Findings

Improves accuracy of pre-trained language models

Outperforms existing intervention baselines

Enhances sample efficiency in activation interventions

Abstract

Modern language models (LMs) have significantly advanced generative modeling in natural language processing (NLP). Despite their success, LMs often struggle with adaptation to new contexts in real-time applications. A promising approach to task adaptation is activation intervention, which steers the LMs' generation process by identifying and manipulating the activations. However, existing interventions are highly dependent on heuristic rules or require many prompt inputs to determine effective interventions. This paper proposes a layer-wise additive activation intervention framework that optimizes the intervention process, thus enhancing the sample efficiency. We benchmark our framework on various datasets, demonstrating improvements in the accuracy of pre-trained LMs and competing intervention baselines.