Effectively Prompting Small-sized Language Models for Cross-lingual Tasks via Winning Tickets

arXiv:2404.01242v1 Announce Type: new
Abstract: Current soft prompt methods yield limited performance when applied to small-sized models (fewer than a billion parameters). Deep prompt-tuning, which entails prepending parameters in each layer for enhanced efficacy, presents a solution for prompting small-sized models, albeit requiring carefully designed implementation. In this paper, we introduce the Lottery Ticket Prompt-learning (LTP) framework that integrates winning tickets with soft prompts. The LTP offers a simpler implementation and requires only a one-time execution. We demonstrate LTP on …

abstract arxiv billion cross-lingual cs.cl current implementation language language models layer paper parameters performance prompt prompting small solution tasks tickets type via