WORKSHOP (INTERNATIONAL) Verbosity Bias in Preference Labeling by Large Language Models

Keita Saito (University of Tsukuba), Akifumi Wachi, Koki Wataoka, Youhei Akimoto (University of Tsukuba)

Workshop on Instruction Tuning and Instruction Following at NeurIPS 2023. (Instruction Workshop @ NeurIPS 2023.)

December 16, 2023

In recent years, Large Language Models (LLMs) have witnessed a remarkable surge in prevalence, altering the landscape of natural language processing and machine learning. One key factor in improving the performance of LLMs is alignment with humans achieved with Reinforcement Learning from Human Feedback (RLHF), as for many LLMs such as GPT-4, Bard, etc. In addition, recent studies are investigating the replacement of human feedback with feedback from other LLMs named Reinforcement Learning from AI Feedback (RLAIF). We examine the biases that come along with evaluating LLMs with other LLMs and take a closer look into verbosity bias – a bias where LLMs sometimes prefer more verbose answers even if they have similar qualities. We see that in our problem setting, GPT-4 prefers longer answers more than humans. We also propose a metric to measure this bias.

Paper : Verbosity Bias in Preference Labeling by Large Language Modelsopen into new tab or window (external link)