Fine-Tuned 'Small' LLMs (Still) Significantly Outperform Zero-Shot Generative AI Models in Text Classification
This addresses the practical need for efficient and accurate text classification tools, showing that fine-tuning remains superior despite the promise of generative AI, which is incremental as it confirms existing trends with new comparisons.
The paper tackled the problem of whether zero-shot generative AI models can replace fine-tuned smaller LLMs for text classification, finding that fine-tuned models consistently and significantly outperform larger zero-shot models across diverse tasks and datasets.
Generative AI offers a simple, prompt-based alternative to fine-tuning smaller BERT-style LLMs for text classification tasks. This promises to eliminate the need for manually labeled training data and task-specific model training. However, it remains an open question whether tools like ChatGPT can deliver on this promise. In this paper, we show that smaller, fine-tuned LLMs (still) consistently and significantly outperform larger, zero-shot prompted models in text classification. We compare three major generative AI models (ChatGPT with GPT-3.5/GPT-4 and Claude Opus) with several fine-tuned LLMs across a diverse set of classification tasks (sentiment, approval/disapproval, emotions, party positions) and text categories (news, tweets, speeches). We find that fine-tuning with application-specific training data achieves superior performance in all cases. To make this approach more accessible to a broader audience, we provide an easy-to-use toolkit alongside this paper. Our toolkit, accompanied by non-technical step-by-step guidance, enables users to select and fine-tune BERT-like LLMs for any classification task with minimal technical and computational effort.