ESG Classification by Implicit Rule Learning via GPT-4

TL;DR

This paper explores how GPT-4 can be guided to evaluate ESG factors effectively without explicit training data, using prompting and reasoning strategies, demonstrating promising results in financial task ranking.

Contribution

It introduces methods to align GPT-4 with unknown ESG criteria through prompting and reasoning, achieving competitive performance without additional training.

Findings

GPT-4 can rank ESG impact types effectively with prompting strategies.

Longer pre-training correlates with better financial task performance.

Prompt adjustments influence model's ability to handle financial evaluations.

Abstract

Environmental, social, and governance (ESG) factors are widely adopted as higher investment return indicators. Accordingly, ongoing efforts are being made to automate ESG evaluation with language models to extract signals from massive web text easily. However, recent approaches suffer from a lack of training data, as rating agencies keep their evaluation metrics confidential. This paper investigates whether state-of-the-art language models like GPT-4 can be guided to align with unknown ESG evaluation criteria through strategies such as prompting, chain-of-thought reasoning, and dynamic in-context learning. We demonstrate the efficacy of these approaches by ranking 2nd in the Shared-Task ML-ESG-3 Impact Type track for Korean without updating the model on the provided training data. We also explore how adjusting prompts impacts the ability of language models to address financial tasks leveraging smaller models with openly available weights. We observe longer general pre-training to correlate with enhanced performance in financial downstream tasks. Our findings showcase the potential of language models to navigate complex, subjective evaluation guidelines despite lacking explicit training examples, revealing opportunities for training-free solutions for financial downstream tasks.