A Dutch Financial Large Language Model

TL;DR

This paper introduces FinGEITje, a Dutch financial LLM, along with a large instruction dataset and evaluation benchmark, demonstrating superior performance in financial tasks with an open-source, automated approach.

Contribution

It presents the first Dutch financial LLM, a large instruction dataset, and an automated evaluation method, advancing multilingual financial NLP capabilities.

Findings

FinGEITje outperforms existing models on five financial tasks.

The dataset and evaluation method are openly available for research.

Automated evaluation reduces manual effort in performance assessment.

Abstract

This paper presents FinGEITje, the first Dutch financial Large Language Model (LLM) specifically designed and optimized for various financial tasks. Together with the model, we release a specialized Dutch financial instruction tuning dataset with over 140,000 samples, constructed employing an automated translation and data processing method. The open-source data construction method is provided, facilitating the creation of financial instruction datasets in different languages. To evaluate model performance, the study introduces the first Dutch financial evaluation benchmark, along with an automated evaluation method that utilizes an LLM as an independent evaluator, reducing manual intervention in performance evaluation. The experimental results highlight the superior performance of FinGEITje across five critical Dutch and English financial tasks.