Enhancing Low-Resource LLMs Classification with PEFT and Synthetic Data
This addresses the problem of computational inefficiency in few-shot LLM classification for low-resource settings, though it appears incremental as it builds on existing PEFT and data generation techniques.
The paper tackles the efficiency-accuracy trade-off in low-resource text classification with LLMs, proposing a method using synthetic data and PEFT to achieve comparable or better accuracy than in-context learning while maintaining 0-shot efficiency, with competitive results on multiple datasets.
Large Language Models (LLMs) operating in 0-shot or few-shot settings achieve competitive results in Text Classification tasks. In-Context Learning (ICL) typically achieves better accuracy than the 0-shot setting, but it pays in terms of efficiency, due to the longer input prompt. In this paper, we propose a strategy to make LLMs as efficient as 0-shot text classifiers, while getting comparable or better accuracy than ICL. Our solution targets the low resource setting, i.e., when only 4 examples per class are available. Using a single LLM and few-shot real data we perform a sequence of generation, filtering and Parameter-Efficient Fine-Tuning steps to create a robust and efficient classifier. Experimental results show that our approach leads to competitive results on multiple text classification datasets.