Grade Score: Quantifying LLM Performance in Option Selection

TL;DR

This paper presents the Grade Score, a new metric for evaluating LLMs' fairness and consistency in multiple-choice tasks, and explores techniques to optimize it, revealing variability among models and the impact of prompt strategies.

Contribution

Introduces the Grade Score metric for assessing LLM fairness and consistency, and demonstrates methods to improve LLM performance using prompt engineering and sampling strategies.

Findings

Grade Score effectively measures order bias and choice stability.

Prompt engineering and sampling strategies improve LLM performance.

Instruction-following models adapt to bias-specific instructions.

Abstract

This study introduces the "Grade Score", a novel metric designed to evaluate the consistency and fairness of Large Language Models (LLMs) when used as multiple-choice judges with respect to order bias and choice consistency. The Grade Score combines Entropy, which measures order bias, and Mode Frequency, which assesses choice stability, offering insights into LLMs' reliability and impartiality. The study explores techniques such as prompt engineering and option sampling strategies to optimize the Grade Score, demonstrating their effectiveness in enhancing LLMs' performance. Results showcase varying performances among LLMs with respect to prompts and highlight the positive impact of including irrelevant options. The study also identifies an emergent behavior in instruction-following models, where they adapt to instructions targeting specific biases, demonstrating their adaptability. The Grade Score facilitates comparisons between LLMs and encourages ongoing research towards optimizing their decision-making processes, with potential implications for improving their reliability and fairness in various applications. All code is available on GitHub https://github.com/IoDmitri/GradeLab