Ultra-Low-Dimensional Prompt Tuning via Random Projection

TL;DR

ULPT is a novel prompt tuning method that drastically reduces training parameters by optimizing prompts in a low-dimensional space with random projection, maintaining high performance across NLP tasks.

Contribution

The paper introduces ULPT, a simple low-dimensional prompt tuning approach that significantly reduces parameters while preserving performance.

Findings

ULPT achieves 98% reduction in training parameters.

ULPT outperforms recent parameter-efficient tuning methods.

ULPT is effective across over 20 NLP tasks.

Abstract

Large language models achieve state-of-the-art performance but are increasingly costly to fine-tune. Prompt tuning is a parameter-efficient fine-tuning method that addresses parameter-efficiency by learning prompt embeddings, but these embeddings are typically tied to the model's hidden dimensionality, limiting parameter saving. In this paper, we propose Ultra-Low-dimensional Prompt Tuning (ULPT), a simple yet effective method that optimizes prompts in a low-dimensional space (e.g., 2D) and uses a frozen random matrix for up-projection. ULPT can achieve 98% reduction in the training parameters compared to vanilla prompt tuning while preserving performance. Our extensive experiments across over 20 NLP tasks demonstrate that ULPT consistently outperforms recent parameter-efficient tuning methods using significantly fewer parameters, making it well-suited as a storage-efficient framework for massive LLM customization.