D\'olares or Dollars? Unraveling the Bilingual Prowess of Financial LLMs Between Spanish and English

TL;DR

This paper introduces Toisón de Oro, a bilingual framework with datasets, models, and benchmarks for Spanish-English financial NLP, demonstrating that specialized bilingual training improves LLM performance in Spanish finance tasks.

Contribution

The paper presents the first comprehensive bilingual financial NLP framework, including datasets, a finetuned LLM, and an evaluation benchmark, addressing the Spanish-English performance gap.

Findings

FinMA-ES outperforms GPT-4 in Spanish financial tasks.

Bilingual instruction tuning enhances LLM performance in Spanish finance.

Existing LLMs show significant multilingual performance gaps and biases.

Abstract

Despite Spanish's pivotal role in the global finance industry, a pronounced gap exists in Spanish financial natural language processing (NLP) and application studies compared to English, especially in the era of large language models (LLMs). To bridge this gap, we unveil Tois\'on de Oro, the first bilingual framework that establishes instruction datasets, finetuned LLMs, and evaluation benchmark for financial LLMs in Spanish joint with English. We construct a rigorously curated bilingual instruction dataset including over 144K Spanish and English samples from 15 datasets covering 7 tasks. Harnessing this, we introduce FinMA-ES, an LLM designed for bilingual financial applications. We evaluate our model and existing LLMs using FLARE-ES, the first comprehensive bilingual evaluation benchmark with 21 datasets covering 9 tasks. The FLARE-ES benchmark results reveal a significant multilingual performance gap and bias in existing LLMs. FinMA-ES models surpass SOTA LLMs such as GPT-4 in Spanish financial tasks, due to strategic instruction tuning and leveraging data from diverse linguistic resources, highlighting the positive impact of cross-linguistic transfer. All our datasets, models, and benchmarks have been released.