Multitask Prompt Tuning Enables Parameter-Efficient Transfer Learning


Published on




Prompt tuning, in which a base pretrained model is adapted to each task via conditioning on learned prompt vectors, has emerged as a promising approach for efficiently adapting large language models to multiple downstream tasks. However, existing methods typically learn soft prompt vectors from scratch, and it has not been clear how to exploit the rich cross-task knowledge with prompt vectors in a multitask learning setting. We propose multitask prompt tuning (MPT), which first learns a single transferable prompt by distilling knowledge from multiple task-specific source prompts. We then learn multiplicative low rank updates to this shared prompt to efficiently adapt it to each downstream target task. Extensive experiments on 23 NLP datasets demonstrate that our proposed approach outperforms the state-of-the-art methods, including the full finetuning baseline in some cases, despite only tuning 0.035% as many task-specific parameters.

Please cite our work using the BibTeX below.

title={Multitask Prompt Tuning Enables Parameter-Efficient Transfer Learning},
author={Zhen Wang and Rameswar Panda and Leonid Karlinsky and Rogerio Feris and Huan Sun and Yoon Kim},
booktitle={The Eleventh International Conference on Learning Representations },
Close Modal