This academic article by Yue Guo, Zian Xu, and Yi Yang from The Hong Kong University of Science and Technology assesses the capability of Large Language Models (LLMs) like ChatGPT in performing financial Natural Language Processing (NLP) tasks. The researchers introduced FinLMEval, a framework for evaluating financial language models. The study compared the performance of encoder-only and decoder-only language models, revealing that while some decoder-only LLMs show decent performance across most financial tasks, they generally fall behind fine-tuned expert models, particularly when dealing with proprietary datasets. The paper aims to provide foundational evaluations for efforts to develop more advanced LLMs in the financial domain.
Publication date: 20 Oct 2023
Project Page: Not Provided
Paper: https://arxiv.org/pdf/2310.12664
