Residual Prompt Tuning: Improving Prompt Tuning with Residual Reparameterization
This addresses the problem of making prompt tuning more effective and robust for users of large language models, representing an incremental improvement over existing methods.
The paper tackled the performance and stability issues of prompt tuning in parameter-efficient tuning of pre-trained language models by introducing Residual Prompt Tuning, which uses residual reparameterization to achieve a +7 point improvement on SuperGLUE with T5-Base and allows a 10x reduction in prompt length without performance loss.
Prompt tuning is one of the successful approaches for parameter-efficient tuning of pre-trained language models. Despite being arguably the most parameter-efficient (tuned soft prompts constitute <0.1% of total parameters), it typically performs worse than other efficient tuning methods and is quite sensitive to hyper-parameters. In this work, we introduce Residual Prompt Tuning - a simple and efficient method that significantly improves the performance and stability of prompt tuning. We propose to reparameterize soft prompt embeddings using a shallow network with a residual connection. Our experiments show that Residual Prompt Tuning significantly outperforms prompt tuning on SuperGLUE benchmark. Notably, our method reaches +7 points improvement over prompt tuning with T5-Base and allows to reduce the prompt length by 10x without hurting performance. In addition, we show that our approach is robust to the choice of learning rate and prompt initialization, and is effective in few-shot settings.