Can domain-specific large language models (LLMs) can outperform general-purpose LLMs in financial prediction tasks?

Description

A comprehensive and empirically rigorous analysis of how domain-specific large language models (LLMs) can outperform general-purpose LLMs in financial prediction tasks, particularly stock return forecasting. Authored by Eghbal Rahimikia of the University of Manchester and Felix Drinkall of the University of Oxford, tackling limitations seen in most current LLM applications in finance, especially the issue of look-ahead bias.