Leveraging Log Probabilities in Language Models to Forecast Future Events

TL;DR

This paper presents a new approach using Large Language Models and log probabilities to improve the accuracy of forecasting future events across various sectors, demonstrating significant performance gains.

Contribution

Introduces a novel multi-step method leveraging LLMs and log probabilities for future event prediction, outperforming existing AI systems.

Findings

Achieved a Brier score of 0.186, indicating high forecast accuracy.

Demonstrated a 26% improvement over random chance.

Achieved a 19% improvement over existing AI systems.

Abstract

In the constantly changing field of data-driven decision making, accurately predicting future events is crucial for strategic planning in various sectors. The emergence of Large Language Models (LLMs) marks a significant advancement in this area, offering advanced tools that utilise extensive text data for prediction. In this industry paper, we introduce a novel method for AI-driven foresight using LLMs. Building on top of previous research, we employ data on current trends and their trajectories for generating forecasts on 15 different topics. Subsequently, we estimate their probabilities via a multi-step approach based on log probabilities. We show we achieve a Brier score of 0.186, meaning a +26% improvement over random chance and a +19% improvement over widely-available AI systems.